NEW U.S. GAS LINES WILL RESTRUCTURE NORTH AMERICAN GRID FLOWS

Eric Spiegel, Elbert Johnson Jr.
Booz-Allen & Hamilton Inc.
Dallas

Albert Viscio
New York

Completion of several major U.S. natural-gas pipeline projects will significantly change relationships among suppliers, buyers, and transporters; alter pipeline flows and tariffs; and affect producer economics.

The competitive and regulatory environment of the natural-gas industry continues to change under great uncertainty.

Within this rapidly changing environment, many longdiscussed but often-delayed pipeline projects are nearing or have entered the construction phase.

These major projects represent more than 5 bcf/day (bcfd) of capacity targeting three major markets that now consume an average of 23 bcfd (Table 1).

Several smaller projects will also be critical to increased flexibility by providing new interconnections with the existing interstate grid for customers in St. Louis, Chicago, Indianapolis, other Midwestern cities, and parts of Virginia.

Various projects in Oklahoma, the Rocky Mountains, New Mexico's San Juan basin, and Michigan will likewise provide greater access to the interstate grid for producers.

In short, the links between supply basins and major markets are increasing dramatically (Fig. 1).

Participants in the projects come from every segment of the industry: producers who wish to increase netbacks and ensure an outlet for transportation-constrained reserves; pipelines that wish to diversify markets, supply existing facilities, and extend their core businesses; and local distribution companies (LDCS) and end users that wish to ensure security of supply.

Producers, who have traditionally stayed out of the interstate pipeline business, stand to gain by expanding their access to Canadian production.

This increased access to markets will translate into increased demand for the region's supply and therefore higher prices for all gas in the area.

Pipeline companies will diversify markets which are increasingly at risk or help ensure that supplies flow into existing systems. For Northeast LDCS, the objective is supply diversity.

Integrated producer, transmission, and distribution companies' intentions could be a mix of any or all of these motives.

An understanding of the competitive dynamics in the new environment of open-access transportation is essential for analyzing changes in utilization along major pipeline corridors as a result of significant capacity additions and shifts in regional supply and demand.

The analysis here draws heavily on a computer model of the North American transportation grid. The model is used under several scenarios to help demonstrate how pipeline transportation flows could be altered.

Finally, there are some key success factors which cannot be quantified in the model.

COMPETITIVE DYNAMICS

Open-access transportation has resulted in much greater competition among interstate pipelines. The level of competition, however, varies significantly from one corridor to the next (Table 2).

Some corridors, such as Permian-California and Midcontinent-Midwest, enjoy very high capacity utilization, and individual pipelines are able to charge the maximum regulated rate in most instances. In the case of Alberta-California, Pacific Gas Transmission (PGT) has a virtual monopoly and runs close to capacity without discounting.

At the other extreme, capacity utilization on the Gulf Coast-Midwest corridor runs as low as 50% on an annual basis for some pipelines. In this case, discounting is the norm.

Whereas regulated rates vary from about 26 to 46/Mcf, actual discounted rates nearly equilibrate, typically 25 to 30 in the winter. On the very large Gulf Coast-Mid-Atlantic-New York/New Jersey-New England corridor, individual pipelines exhibit different pricing strategies.

Tennessee Gas Pipeline Co. (Tennessee) and Columbia Gas Transmission Corp. (Columbia) take advantage of relatively low costs and capture transportation volumes without significant discounting. Transcontinental Gas Pipe Line Corp. (Transco) counters higher costs with more discounting but also relies on LDC customers, who have converted sales volumes to firm transportation, to discount as well.

Texas Eastern Transmission Corp. (Texas Eastern), also higher cost, retains a much greater share of traditional sales throughput. Under traditional utility regulation, lost throughput historically led to increased regulated rates to allow pipelines to continue to meet "revenue requirements."

In today's competitive environment, the response to lost throughput is usually discounting, in some cases well below the regulated rate, to avoid a "death spiral."

Throughout the 1980s, pipeline companies competed for additional throughput primarily by installing incremental compression, interconnecting with other existing pipelines, making greater use of existing storage, and of course by discounting rates.

The impact of their investments (i.e., excluding rate adjustments) was not very dramatic. The proposed new pipelines and major expansions, however, will restructure flows.

Some existing pipelines will lose throughput and therefore be pressured to adjust rates. Other corridors will look more like today's Gulf Coast-Midwest corridor.

On the surface, the rationale for some of the projects appears straightforward. For example, the Kern River and WyCal projects can point to a price differential between the Rockies (delivered to pipeline) and the California border of 70 to $1/Mcf compared to estimated tariffs on their new pipelines of only 30 to 50.

Projects such as the Natural Gas Pipeline Co. of America (NGPL) interconnect and Arkla Inc.'s AC line out of the Midcontinent can point to the rapid increase in Arkoma basin (Oklahoma) deliverability and the high capacity utilization on existing pipelines out of Oklahoma.

But, the impact of these projects extends beyond their individual corridors.

The North American pipeline grid is extremely complex. The interconnectivity of individual pipelines means that an expansion or rate change can have a ripple effect thousands of miles away.

Moreover, regional supply and demand are constantly changing, adding uncertainty to the equation. The uncertainty of future consumption can be illustrated by comparing the most optimistic and most pessimistic "base cases" from three commonly used, publicly available forecasts.

The Gulf Coast, where many expect onshore production to continue its historical decline, exemplifies the uncertainty of regional supply. Pipelines out of the Gulf Coast must be concerned about the ability of now maturing offshore production to continue partially to offset the onshore decline (Fig. 2).

THROUGHPUT IMPLICATIONS

To help in understanding the impact of the new pipelines and major expansions, as well as shifts in regional supply and demand, a computer model of the North American transportation grid was developed.

An optimization algorithm (similar to linear programming) allocates the lowest delivered-cost gas to its most highly valued use. Demand is modeled separately for 13 geographic regions, each with two sectors-a highly price-elastic fuel-switchable sector and a relatively price-inelastic nonswitchable sector.

Alternative fuel prices also vary by demand region. Supply is modeled separately for 13 geographic regions, each with an assumed level of deliverability and minimum or "shut-in" wellhead price.

The supply and demand regions are linked by a transportation grid reflecting the capacity and typical transportation rates along individual pipeline corridors. Table 3 summarizes the assumptions for pipeline capacity additions.

Notice that not all of the projects are assumed to be built. For example, only one of either Kern River or WyCal is programmed into the model. Similarly, only one of either the El Paso Natural Gas Co. (EPNG) or Transwestern Pipeline Co. expansions is assumed to be constructed.

The proposed PGT expansion from Canada to California is also included in the future grid but not the proposed Altamont system which would link Canadian supplies to both California and the Midwest. Instead of providing a single outlook, we demonstrate a range of plausible outcomes for pipeline throughput via a series of four supply cases.

These cases, based on discussions and published industry analyses, assume the following:

No change from current deliverability (production plus bubble).
Decline-Long-term historical decline rate for Permian (5.4%/year) and Gulf Coast onshore (5.3%/year); all other supply regions unchanged from current deliverability.
Growth-Additional 1 bcfd deliverability for Midcontinent and Rocky Mountain supply regions; all other supply regions unchanged from current deliverability.
Shift-Long-term historical decline rate for Permian (5.4%/year) and Gulf Coast onshore (5.3%/year); additional 1 bcfd deliverability for Midcontinent and Rocky Mountain supply regions; all other supply regions unchanged from current deliverability.

We have assumed moderate growth in demand overall (more in some areas such as the Northeast and less in others) of approximately 3%/year for switchable customers and 1 %/year for nonswitchable customers, all else being equal.

Consumption, prices, and pipeline throughput are determined by the interaction of regional demand "curves," alternate fuel prices, and the various supply assumptions.

These scenarios bracket a range of potential outcomes, summarized by major corridors in Table 4 for the Northeast, Midwest, and California markets. The data reflect modeled conditions for the year 2000.

In each exhibit, the predicted capacity utilization rate, as portion of maximum practical annual capacity considering seasonal and other effects, is provided for each supply and demand scenario.

Lower utilization rates imply greater competition and increased discounting. An interpretation of the quantitative results is provided in the following discussion.

NORTHEAST

The Gulf Coast will continue to be the primary supplier to the Northeast via the existing Gulf Coast-Mid-Atlantic corridor. Capacity utilization on the corridor, however, is susceptible to declining reserves and production on the Gulf Coast and to growing demand on the Gulf Coast and in the Southeast.

(Table 4 shows utilization on the corridor beyond the Southeast; short-haul throughput is greater.)

For example, the pessimistic supply scenario results in a throughput loss of up to 40% of the corridor's capacity. Permian supplies backed out of California by new projects as well as Midcontinent supplies will partially offset any decline in Gulf Coast production.

Midcontinent supplies will gain greater access to growing Northeast markets along two corridors, even though the transportation costs are higher than those from the Gulf Coast.

Gas can flow along a "northern route" to the Midwest on existing systems (Panhandle Eastern Pipe Line Co., ANR Pipeline Co., or NGPL, for example) and then onto new and expanded lines (such as Panhandle, ANR, CNG Transmission Corp., or Texas Eastern) which comprise an interconnect to the Mid-Atlantic. Or, gas can flow along a "southern route" via mostly new lines (Arkla AC, NGPL interconnect, or Arkansas-Oklahoma) which comprise an interconnect to the existing Gulf Coast-Mid-Atlantic corridor discussed previously (Texas Eastern, Tennessee, Transco, Columbia Gulf, or Texas Gas Transmission Corp.).

Combined capacity on the two interconnects will exceed throughput under all scenarios, resulting in intense rate competition. We have assumed this rate competition will lead to similar capacity utilization for modeling purposes.

Note, however, that capacity-utilization increases significantly on the two interconnects if Midcontinent supplies grow or Gulf Coast supplies decline. Declining Gulf Coast supplies would also lower utilization into New York/New Jersey and New England, but these corridors will capture Midcontinent supplies whether they take the northern route or the southern route.

Canadian gas will also find its way to Northeast U.S. markets via Iroquois Gas Pipeline or other new projects, largely because pipeline capacity to California will be fully utilized. Assuming a single competitive market price in the Northeast, netbacks to Canadian producers from California (or the Midwest) would exceed those from the Northeast by as much as 85 Mcf.

Iroquois will achieve higher-than-predicted utilization, however, if Canadian suppliers can successfully sell "supply diversity" (which is not captured by the model) and realize somewhat higher prices in the marketplace than do U.S. suppliers.

In summary, new projects and shifting regional supplies will increase competition for existing pipelines into the Northeast. The new or expanded pipelines will also face intense competition because pipeline-capacity additions will exceed demand growth (at least in the near-term), and the target customers' choices will widen to include suppliers in Canada and the Midcontinent.

Existing pipelines out of the Gulf Coast can offset potentially declining Gulf Coast supplies with Permian supplies via existing Texas intrastates and Midcontinent supplies via the new Midcontinent-Gulf Coast interconnect.

These corridors will link up to compete with the northern route from the Midcontinent (which includes several new projects) as well as new pipelines from Canada which capture the "overflow" of Canadian production not bound for eastern Canada, the U.S. Midwest, or California.

This competition is summarized quantitatively in Table 4, which lists predicted utilization rates for each corridor under the various scenarios.

MIDWEST

The existing Rockies-Midcontinent corridor could lose significant throughput when the new Rockies-California corridor draws its traditional supplies westward. The damage will be less if Rockies' supplies grow (as indicated by the model results) or if Altamont brings Canadian gas into the area (not modeled).

The combination of this lost flow into the Midcontinent and the outward flow of local supplies bound for the Northeast along the Midcontinent Gulf Coast interconnect (discussed previously), will decrease supplies available to the existing Midcontinent-Midwest corridor.

Pipelines on this major corridor, which currently enjoy very high utilization, can be expected to lose throughput and begin discounting. These losses would be dampened by growing supplies in the Rockies or Midcontinent.

Results for the Gulf Coast-Midwest and Alberta-Midwest corridors appear to run counter to what might otherwise be expected: Utilization decreases under the supply "growth" scenario and increases under the supply "decline" scenario.

The explanation lies in the pipeline grid's flexibility to respond to changing conditions by, for example, redistributing incremental supplies from just two areas (the Rockies and the Midcontinent under the supply "growth" scenario) so that all market areas increase consumption.

Under the supply "growth" scenario, California's incremental supply can only come from the Permian basin because pipelines from Canada and the Rockies are full (all San Juan gas being already bound for there within the model) under the "no change" case.

This in turn results in less Permian gas moving east into Texas. Moreover, the Gulf Coast market area's incremental supply can only come from the Gulf Coast supply area. The net result is less excess supply on the Gulf Coast available for export to Midwest markets.

Similarly, Alberta supplies "stay home" to allow increased consumption in both eastern and western Canadian markets.

The deficit for the Midwest is made up from the Rockies and the Midcontinent (where supplies are assumed to grow) along previously underutilized corridors.

In summary, pipelines serving the Midwest will experience increased competition for two reasons.

First, Canadian supplies will gain greater access to the market, backing out U.S. supplies.

Second, Rocky Mountain and Midcontinent supplies traditionally locked into Midwest markets will gain greater access to California and the Northeast via new projects and therefore be drawn away from pipelines into the Midwest.

This competition is summarized in Table 4.

CALIFORNIA

Over the past 2 years, industry experts and trade publications have focused much of their attention on the large incremental gas market in California.

Although market-growth projections remain strong, the consensus of what drives the need for new capacity has changed as the enhanced oil recovery (EOR) market faded in the eyes of producers and pipelines and has been replaced by a more positive outlook in power markets. In addition, current proposed capacity still greatly exceeds all forecast requirements, and the presence of too many projects has hindered expansions to date.

The transportation model employed in this study shows that the addition of major pipelines in the Alberta-California and Rockies-California corridors could significantly alter North American gas flows (OGJ, Oct. 1, p. 79; Oct. 8, p. 110).

As a result, the California market will see the most dramatic shift in supply over the next decade.

Under all scenarios the new and expanded corridors (from Alberta and the Rockies) would be fully utilized, while the Permian-San Juan and San Juan-California corridors would see a decline in utilization from current levels under most scenarios.

Capacity utilization on those corridors is highly sensitive to declining reserves and production in the Permian and additional deliverability in the Rockies and Midcontinent. For example, the "decline" and "shift" supply scenarios result in a 40-50% throughput loss from the Permian to San Juan and a 20-30% loss from San Juan to California (Table 4).

These results demonstrate that the new pipelines would actually serve to displace existing pipeline throughput as well as meet the growing market in California. Existing contracts with EPNG and Transwestern expire in 1991 and 1994, respectively, easing the possible displacement.

While the model does not include an Alberta-to-Rockies pipeline (Altamont), this corridor would have the advantage of moving gas west to California or east to the Midwest.

In summary, pipelines serving California will not only experience competition for the large forecast incremental demand but may struggle to maintain throughput on some existing capacity. As previously noted, the diversion of Canadian and Rockies supplies to California could have a negative impact on northern corridors to the Midwest and Northeast.

SUCCESS FACTORS

The analysis verifies that competition will increase among corridors, reaching some that so far have been relatively insulated. Although not addressed explicitly, competition will therefore increase within corridors as well, that is, among individual pipelines linking the same supply basins and markets.

The most important message for the new projects is that high capacity utilization cannot be guaranteed. As a result, the new pipelines may be forced to discount below regulated rates and accept less than "normal" returns on investment.

Even before they are constructed, the new projects are competing on several key success factors: cost, mix of partners, contract commitments, operational factors, and impact on existing customers.

COST

The increasing competition and commodity nature of pipeline transportation imply that low costs will be the primary key to success. Because customers will have a wide range of suppliers, of both natural gas and alternative fuels, prices will generally equilibrate in a given market.

Likewise, producers will normally have a choice of outbound pipelines (issues of receipt point flexibility not withstanding), forcing all transporters to offer competitive rates. With similar pricing, only the lowest cost transporters will be profitable.

PARTNERS

The mix of partners is a key risk-reducing vehicle. Almost all projects have some blend of buyer, seller, and transporter. The advantage is that each partner has a stake in some other part of the value chain, thus bringing expertise to the project.

Moreover, equity owners are also shippers, suppliers, or buyers on the new system and can influence the economic outlook as well as help accelerate project development. Most projects show a significant level of support from either producers or purchasers, but few have enough at both ends of the pipe to be completely balanced.

CONTRACTS

The participants also mitigate risk by lining up shippers in advance. This approach is similar to new shopping malls or office buildings first securing a major "anchor tenant" to ensure sufficient scale to be competitive. Other, smaller "tenants" can then be added at low marginal cost.

Projects try to "fully subscribe" their proposed capacity with firm customers (Table 5). Producers, end users, LDCS, or other pipelines commit (to varying degrees of certainty) to long-term contracts for firm transportation service. Risk is transferred to these customers when they agree to pay demand charges that are independent of the volumes they ultimately ship on the pipeline.

But the demand charges involved represent only a modest portion of the revenues the projects will need to be profitable. The degree to which the contracts guarantee profitability in advance is a function of the rate structure and the length of the contracts.

The primary issue is recovering the full investment, not just the variable cost. For example, assuming half of the transportation tariff is comprised of a demand charge and half of a commodity charge, and that the investment will be recovered over a 25-year period, a pipeline fully subscribed with 10-year contracts has guaranteed roughly 20% of the revenue required to earn the regulated maximum return.

And while some contracts extend for as long as 15 years, most are for 5 or 10, and some are for as short as 2 years.

In addition to volume, the type of shippers who are committing is also important. The Champlain project may have lost its bid to serve New England in part because of an over-reliance on electric generators and cogenerators, whose future requirements are less certain than those of Iroquois' LDC customers.

Kern River may be similarly at a disadvantage relative to WyCal because of a higher mix of EOR producers.

OPERATIONAL FACTORS

Nearly all the projects claim to access "low cost" or "secure" supplies, but some also offer unique operational advantages.

For example, NGPL stresses the added "flexibility" its system now affords Midcontinent producers by directly accessing Midwest markets on either the Amarillo or Gulf Coast legs and indirectly accessing California (as before) and now Northeast markets.

Oklahoma-Arkansas (OGJ, Nov. 6, 1989, p. 25) points to a "unique option" of multiplerate structures.

The Arkla/ANR/Texas Gas project (line AC which began flowing gas Nov. 1), on the other hand, stresses access to either Arkoma or traditional Midcontinent production and a "three-in-one" concept whereby each of the project owners controls a share of the new capacity.

Access to the new line really means access to three separate pipeline systems, each with its own tariff structure and markets in the Midwest and Northeast. The shipper can take his or her pick and communicate only with the most attractive one.

The ability to get pipe into the ground first is another important operational consideration. The ANR/Panhandle Eastern and Texas Eastern/CNG projects into the Northeast are already under construction, unlike the long-delayed Iroquois and brand new Empire State projects.

The NGPL and Arkla/ANR /Texas Gas projects have the same advantage over other Midcontinent proposals. The industry is still anxiously waiting to see who goes first in the race to California.

EXISTING CUSTOMERS

Finally, regulatory treatment and the price responsiveness of existing customers will also affect the new projects' abilities to compete first for the right to install new capacity and then the throughput. Arkla and El Paso create an advantage by 11 rolling in" the new investment into the existing rate base and lowering rates for everyone already on the system.

Others, such as PGT, will be required to treat the new investment as "incremental" and develop a rate structure solely for the expansion, thus insulating existing customers.

TransCanada PipeLines has run into trouble with existing customers who would suffer from the proposed rolled-in approach. This expansion and the associated downstream Iroquois project are now at risk because incremental treatment would require significantly higher tariffs from new customers, making it much more difficult to compete with corridors from the Midcontinent and Gulf Coast.

CHANGES

Major new pipeline projects with more than 6 bcfd of capacity represent a powerful new force of change for U.S. natural gas suppliers, transporters, and users.

The level of competition on existing pipeline corridors today varies. Several corridors enjoy high capacity utilization that allows individual pipelines within those corridors to charge regulated transportation rates and enjoy relatively attractive financial returns.

Too much new capacity is scheduled to come on stream, however, to be filled by incremental demand, at least for the next several years. As a result, flows on existing corridors will be restructured as the market adjusts to the new capacity and to shifts in regional supply and demand.

Suppliers and buyers will have greater choice, but transporters will face more intense competition.

Corridors that have traditionally been fully utilized and relatively insulated from competitive pressures will look more like today's Gulf Coast-Midwest corridor where discounting is common.

Outlooks for the new pipeline projects themselves vary. Although they have taken steps to reduce risks, such as joint venturing with different kinds of partners and lining up shipper commitments in advance, they must still achieve high utilization to ensure healthy returns on their investments.

Projects proposed to serve California appear to be the best positioned, assuming they are not all constructed.

Projects targeting the Northeast face more uncertain demand growth. Those projects moving gas out of Canada must overcome higher costs with a hard sell of "supply diversity."

Despite their cost advantage, those projects moving gas from the Midcontinent will still experience intense competition to fill capacity in excess of market requirements.