U.S. Shale Boom Finally Runs Out of Gas

The correction in U.S. natural gas production has finally arrived. Volumes are down close to 5% since peaking in late 2015, driven by declines in associated gas and higher-cost legacy areas as well as ongoing curtailments and pipeline constraints across the Northeast United States. Natural gas prices have rallied as a result, with 2017 Henry Hub futures increasing almost 15%, to an average of $3 per thousand cubic feet.

With record power burn set to moderate, however, and as new pipelines are put into service across the Northeast, tight supply conditions should ease. By 2018, we expect prices to once again be governed by the considerable productive capacity of low-cost shale plays like the Marcellus, Utica, and Haynesville. Long-term demand tailwinds still appear favorable, although we continue to believe that at our midcycle prices of $3/mcf Henry Hub and $55/barrel West Texas Intermediate, there is more than enough economically profitable resource to meet growing consumption needs. The temporary spike in natural gas prices appears to be just that: temporary. Our long-term price assumption of $3/mcf is unchanged.

Dry gas volumes peaked in late 2015 at 75 billion cubic feet per day and have since dropped more than 3.5 bcf/d. The sharp pullback in drilling activity in oil-focused plays such as the Bakken, Eagle Ford, and Permian--where rig counts have fallen more than 80% in some cases--has driven associated gas volumes down 800 million cubic feet per day since their peak in mid-2015. Higher-cost shale plays such as the Barnett and Fayetteville as well as legacy conventional areas have struggled to compete with lower-cost regions like the Marcellus and Utica; as a result, these higher-cost areas have shed close to 6 bcf/d collectively since the start of 2014. The saving grace for U.S. gas production has been the Northeast U.S. and the Haynesville, which in aggregate have held steady over the past several quarters.

As Expected, Near-Term Prices Jumped on Tightening Supply
We made the case in April that a combination of high power burn and ongoing pipeline bottlenecks in the Marcellus and Utica could drive near-term natural gas prices higher than otherwise indicated by the futures curve, despite the high storage levels that existed at the time because of the third-warmest winter on record. Our expectations weren’t too far off the mark, with near-term Henry Hub prices rallying close to $1/mcf (up almost 50%) over the past several months, driven by lower-than-normal injections. Long-term futures prices haven’t meaningfully changed, however.

After updating our supply and demand projections, it appears that natural gas futures prices through the end of 2017 are more or less in line with what we’d expect, based on the historical relationship between relative storage levels and prices. With the U.S. having ended the 2016 injection season at near-record working gas levels of 4 trillion cubic feet, winter weather will have to cooperate to avoid a collapse in near-term prices. On the flip side, if production--especially in the Northeast U.S., which is one of the few areas we expect to grow over the next few quarters--is unable to respond to demand, or if winter weather is colder than expected, near-term prices could spike. Over the next several months, we expect higher-than-normal withdrawals and lower-than-normal injections (driven by declining supply) will help bring down currently elevated working gas inventories.

Infrastructure Constraints Continue to Hold Back Production
Volumes in the Marcellus and Utica have recently flattened, driven by a shortage of pipeline capacity across the Northeast U.S. In addition to limiting new drilling activity, infrastructure constraints have led to significant curtailments over the past several months as operators choke back existing production due to unfavorable regional prices.

Since the start of 2015, the top 10 operators (which collectively account for around 75% of statewide volumes) have curtailed approximately 2 bcf/d of natural gas production in the Pennsylvania Marcellus, with curtailments tending to move in line with regional prices. The implication: Significant volumes can be brought on line once new pipelines and/or an improvement in prices come about.

Since April, Northeast U.S. pipeline projects totaling more than 5 bcf/d of capacity have had their in-service dates pushed back between 6 and 24 months or canceled altogether, driven largely by delays in permitting and/or a lack of customer demand. As a result, we have tempered our near-term Marcellus and Utica production forecasts, with volumes now ramping gradually throughout 2017 as new capacity is added. As long as the Marcellus and Utica are short pipeline capacity, there will is likely to be continued pressure on prices in this region. Beyond 2017, however, there should be sufficient capacity to move volumes to market without a significant haircut to prices. We note, however, that regulatory delays remain a risk to watch.

As Infrastructure Catches Up, Marcellus and Utica Set to Grow Again
We developed our Three P’s framework--productivity, pipelines, and potential resource--as a way to assess how big the Marcellus and Utica could ultimately get (and by extension, how likely they are to crowd out natural gas supply from other regions). Despite the potential for transitory pipeline delays, the other elements of our framework support our thesis on the productive capacity of the Northeast U.S.: (1) There is sufficient low-cost resource in place across the Marcellus and Utica to meet 10-15 years of U.S. natural gas consumption (assuming all demand is met with Northeast U.S. supply) at $3 per mcf or less, and (2) the vast extent of these plays implies a relatively flat U.S. natural gas cost curve.

The Marcellus and Utica continue to set new high-water marks regarding initial production rates. Notably, the declines are shallow, with first-year decreases of approximately 30% compared with 70%-plus across other shale plays like the Haynesville. The combination of high IP rates and shallow declines has helped the Marcellus and Utica to continue increasing production volumes even as rig counts have fallen precipitously.

In our base-case production forecast for the Northeast U.S., we expect dry gas volumes in the Marcellus to increase from 17 bcf/d at year-end 2015 to more than 27 bcf/d by year-end 2020, with the Utica going from around 2.4 bcf/d to 4.5 bcf/d over the same period. Combined, we expect the Marcellus and Utica to deliver an incremental 13 bcf/d of dry gas volumes through the end of the decade, growing at an 11% compound annual rate from 2015 to 2020.

Given the high number of drilled but uncompleted wells, or DUCs, across the Marcellus and Utica--along with the increasing productivity of new wells--rig activity in the Northeast U.S. is unlikely to ramp back up to historical levels (the Marcellus peaked at 140 rigs; the Utica at 45 or so). Still, we believe the Marcellus and Utica can reach a combined dry gas production level of more than 32 bcf/d by the end of 2020, averaging just 45 rigs for the next five years (thanks, in part, to the 1,200 or so DUCs we project will be brought on line through mid-2019).

Revised U.S. Supply Forecast Envisions More Prominent Role for Haynesville
Alongside the Marcellus and Utica, we expect associated gas volumes to be a significant contributor to production growth in the years ahead. In addition, we believe the Haynesville Shale could play a more prominent role in the U.S. supply stack versus our previous forecast, when evaluated using our Three P’s framework, given (1) significant increases in well productivity that have made this region economically competitive with the Northeast U.S.; (2) proximity to key end markets, which reduces transportation costs; (3) sufficient infrastructure to accommodate increasing production volumes; and (4) plenty of drilling inventory to work through over the next several years. With ongoing pipeline delays across the Northeast increasing the risk profile of the Marcellus and Utica, we suspect operators are looking more favorably on low-cost areas like the Haynesville that are long pipeline.

In our forecast for dry natural gas production in the U.S. through 2020, we project the Marcellus will grow by almost 11 bcf/d, followed by associated gas (more than 4 bcf/d), the Haynesville (3 bcf/d), and the Utica (more than 2 bcf/d). Higher-cost shale areas and legacy conventional production should continue to decline throughout our forecast period. In aggregate, we project dry gas volumes will grow at a 3% annual clip, reaching 87 bcf/d by 2020.

Long-Term Demand Forecast Largely Unchanged;
So Is Our Midcycle Natural Gas Price Forecast
We continue to expect long-term U.S. natural gas demand will be driven by pipeline and liquefied natural gas exports as well as modest growth in industrial consumption; our multiyear projections for these buckets remain largely unchanged. In the short term, however, we’ve ratcheted back our power generation demand forecast because of higher natural gas prices, which is likely to limit coal-to-gas switching. By 2020, our revised power gen demand projections converge with our previous forecast. Compared with our previous forecast--and despite lower electric power generation consumption next year--ongoing infrastructure constraints across the Northeast U.S. will probably require higher Canadian imports and working gas withdrawals in 2017 to ensure supply can meet demand.

We’ve seen no evidence that would lead us to revise our expectation for a midcycle natural gas price of $3/mcf. Our midcycle gas price forecast is based on an in-depth analysis of the competitive dynamics between upstream and services firms, an examination of well-level returns for each of the major producing areas in the U.S., and a comprehensive review of the productive potential of the Marcellus, Utica, and Haynesville shales.

Alongside significant growth in associated gas volumes at our midcycle oil price of $55/bbl (WTI), we believe that at $3/mcf, there is sufficient economic incentive for low-cost natural gas producers to ramp up activity to meet demand. Surveying the gas-focused exploration and production firm we cover--particularly those in the low-cost Marcellus, Utica, and Haynesville--we find that at $3/mcf Henry Hub, a number of these firms are able to generate high-single-digit production growth while remaining approximately free-cash-flow-neutral (if not positive) and in good shape from a balance sheet perspective. What this tells us is that if pipeline capacity is able to keep pace (and we believe it will), and if the low-cost inventory exists (and we believe it does), anything above $3/mcf should lead to an acceleration in supply growth, which would in turn lead to lower prices (assuming demand remains unchanged). In other words, the Marcellus, Utica, and Haynesville should serve as the production accelerators and the price brakes for the U.S. natural gas complex.

Also, pipeline capacity across key producing areas like the Marcellus, Utica, and Haynesville should more than keep pace with increasing activity over the medium term, providing additional room to increase production should conditions warrant. Curtailed volumes could provide additional upside to our forecast from these regions.

Finally, we expect additional pressure on services prices in the years ahead, given lower drilling and completion activity required to bring on new supply. All else equal, this dynamic should lower the cost of production across the U.S. oil and gas complex.