Strong Natural Gas Prices And Tight Supply In 2017

Posted in The Petroleum Truth Report on May 16, 2017

A year ago, most analysts were bearish about natural gas prices.   I wrote that natural gas prices might double and they did. Today, most analysts are again bearish about gas prices and again, I think that they are probably wrong at least for 2017.

The mainstream narrative is that new pipeline capacity—notably the Rover Pipeline—out of the Marcellus and Utica shale plays will unleash a torrent of pent-up supply. That is because over-production in these plays has saturated the northeastern U.S. markets and 2016 wellhead prices averaged about $0.88/mmBtu less than Henry Hub prices (Figure 1). New take-away capacity to higher-priced markets will fix that problem but gas prices will plummet later in 2017 because of increased output.

Figure 1. Marcellus Wellhead Prices Were $0.88 per mmBtu Less Than Henry Hub Prices in 2016. Source: MarcellusGas.Org, EIA and Labyrinth Consulting Services, Inc.

Systematic overproduction turned the northeastern U.S. from the highest-margin market to the lowest by 2013. With a second chance to at least be on par with national pricing, shale gas companies will, according to the narrative, over-produce the entire U.S. market to a loss once again. Smart.

Conventional Gas, Shale Gas and Net Imports

There are three components to gas supply: conventional gas production, shale gas production, and imports. These must be understood to establish a context for a potential supply increase from the Marcellus and Utica shale plays.

There is no doubt that low prices resulted in a 4.26 bcf/d (billion cubic feet of gas per day) decline in gas production from September 2015 through October 2016 (Figure 2).

Figure 2. U.S. Gas Production Fell 4.26 bcf/d From September 2015 to October 2016. Source EIA Natural Gas Monthly and Weekly Updates, and Labyrinth Consulting Services, Inc.

Since 2008, conventional gas production has been in terminal decline and has fallen 26 bcf/d. It is currently falling about 3 bcf/d each year. Shale gas–including associated gas from tight oil—now makes up more than two-thirds of domestic supply. That means that shale gas output must grow by more than 3 bcf/d each year to offset falling conventional supply.

But annual shale gas production growth slowed from almost 7 bcf/d in the first quarter of 2015 to less than 2 bcf/d in the first quarter of 2017 (Figure 3).

Figure 3. Shale Gas Growth Has Slowed from Almost 7 bcf/d in the First Quarter of 2015 to Less Than 2 bcf/d in the First Quarter of 2017. Source: EIA Natural Gas Weekly Update and Labyrinth Consulting Services, Inc.

If shale gas production growth doubles in 2017, then supply will be flat but considerably lower than 2015 levels when over-supply crushed gas prices. Gas supply must increase well beyond what is likely this year in order for prices to fall much below current levels of about $3.25 per mmBtu.

Considerable supply potential exists. The shale gas horizontal rig count has more than doubled—from 76 to 167 rigs—since June 2016 with higher gas prices (Figure 4). How quickly can that potential be converted into supply?

Figure 4. Shale Gas Rig Count Has More Than Doubled Since June 2016 With Higher Gas Prices. Source: EIA and Labyrinth Consulting Services, Inc.

EIA’s latest production forecast suggests that it may happen very quickly. The May STEO projects gas growth of 5.6 bcf/d in 2017 which includes an additional 3.5 bcf/d between April and the end of the year (Figure 5).

Figure 5. EIA Forecast is for a 5.6 Bcf/d Gas Production Increase in 2017 with Prices Rising to $3.43 By December. Source: EIA May 2017 STEO and Labyrinth Consulting Services, Inc.

Although that may be unreasonably aggressive, it is noteworthy that the overall supply balance (red and blue fill in the figure) remains in deficit for most of the year, and that spot prices continue to increase, ending the year at almost $3.50/mmBtu. Net imports (the third component of total supply in addition to shale gas and conventional gas) are forecast to average -0.3 bcf/d in 2017 compared to +1.7 bcf/d in 2016.

Rover Pipeline

The Rover Pipeline was certificated for construction in mid-February and will connect gas from the Utica and Southwestern Marcellus shale plays to the Defiance Hub in northwestern Ohio (Figure 6). There is a gas surplus (~1.8 bcf/d) in Ohio so this pipeline is a gas exit route to the Dawn Hub in Ontario, and to the Midwest and Gulf Coast via interconnecting Vector, Panhandle Eastern and ANR pipelines. There, it will compete with existing supply and result in lower prices.

Figure 6. Rover Pipeline Route Connecting Utica and Southwestern Marcellus Shale Plays With the Defiance Hub. Source: Energy Transfer and Labyrinth Consulting Services, Inc.

Although Rover is scheduled to reach Defiance in November, it is unlikely that any gas will move beyond there before 2018. It will not, therefore, have any effect on gas supply in 2017. Depending on how much gas ultimately is sent to Canada, it may have limited effect on U.S. supply in 2018.

What Could Go Wrong?

The consensus of experts has been consistently wrong about natural gas supply for decades. That’s why LNG import terminals were built following gas shortages in the 1970s only to be shuttered after imports from Canada, fuel switching to coal and nuclear, and gas industry deregulation resulted in 15 years of stable gas supply.

By the early 2000s, import terminals were re-opened as Canadian gas production began to decline and domestic output failed to rally even with much higher gas-directed rig counts. The shale revolution ended all of that and now, those import terminals are being re-designed to export LNG. Gas export will likely prove to be fully out-of-phase with future gas supply once again.

That is why I am skeptical when experts now declare an impending gas over-supply. Gas prices remain well above $3/mmBtu after one of the warmest winters on record, and most data suggests that supply will remain tight at least through the end of 2017.

What could go wrong with that hypothesis? Weather, of course, and Morgan Stanley has astutely pointed out that 2016 rainfall in California may displace some natural gas with hydro for electric power generation. They and PointLogic note that some cooler summer forecasts might further reduce gas demand.

At the same time, EIA expects higher-than-average consumption for Summer 2017 (Figure 7) and the Browning World Climate Bulletin predicts a warmer-than-average summer with early El Niño onset.

Figure 7. EIA Forecasts Higher-Than-Average Consumption for Summer 2017. Source: EIA May 2017 STEO and Labyrinth Consulting Services, Inc

Morgan Stanley supposes that associated gas from tight oil plays will be a major factor in increased gas supply. This ignores the considerable  dysfunction in the pressure pumping business where frack crews commonly lag demand by at least 6 months. Rig count increases will probably not translate into production gains as quickly as many oil-price bears assume. Gas pipelines out of the Permian basin remain problematic and most gas from the Eagle Ford will go to Mexico.

Morgan Stanley’s belief that significant expansion of production in the Haynesville Shale will occur is based on incorrect sub-$3.00 break-even prices. Exco–the second largest Haynesville producer–shows a maintenance spending level of about $3.50 in their 2016 10-K after writing off all proved undeveloped reserves in accordance with the SEC 5-year rule.

It also seems unlikely that losses in major gas-producing areas including Texas, Oklahoma, Wyoming, Arkansas, Utah, Louisiana and the OCS Gulf of Mexico will be quickly offset by gains in Ohio, Pennsylvania and West Virginia especially considering frack crew availability (Figure 8).

Figure 8. Unlikely That OH, PA & WV Gains Will Offset TX, OK, WY, AR & OCS Losses in 2017. Source: EIA and Labyrinth Consulting Services, Inc.

Comparative inventories indicate that the mid-cycle price trend has moved upward from $3.00 to $3.60 (or higher) since mid-March reflecting market perception of tight supply (Figure 9). The mid-cycle price—where the trend line intersects the y-axis—represents the median price that the market deems necessary to maintain supply throughout the present price cycle. If this trend persists, it is possible that year-end gas prices will be in the $3.50 to $4.00 range.

Figure 9. Gas Mid-Cycle Price Has Shifted To $3.60/mmBtu or Higher. Black arrows show progression from higher to lower price trend and back again. Source: EIA and Labyrinth Consulting Services, Inc.

At the same time, it is likely that prices will be substantially lower in 2018 once the Rover and other pipelines are operating and frack crews begin catching up with drilling levels. That possibility is reflected in inverted natural gas forward curves (Figure 10). Note that the price for futures contracts drops sharply in January 2017.

Figure 10. Henry Hub Forward Curves Are Inverted and Rising. Source: CME and Labyrinth Consulting Services, Inc.

Although forward curves should never be viewed as a price forecast, they reflect current market expectations. Those expectations seem clear and are supported by all available data: natural gas supply should remain fairly tight through 2017 and will probably increase some time in 2018 and that will result in lower gas prices. Understand the uncertainties and plan accordingly.


28 comments on this entry

  1. I was wondering why natural gas stocks have been performing so poorly relative to the underlying commodity. Usually, thats not a good sign for the commodity. Thanks for you post Art!

  2. Vince,

    Almost all energy stocks have been substantially devalued in the last few months. Some of that is an adjustment from exaggerated valuations and some of it is failing confidence in oil-price recovery. But your point that gas companies have been punished more than oil-weighted companies is valid. That is because of the expectation that new pipelines will bring overwhelming supply and lower prices to the market. There are a few exceptions like Rice and Cabot.

    All the best,


  3. Good points Art. Although I think what the market (by which I mean the forward curve) is telling us that prices spike to the mid $3’s this winter, fall to just above $3 in 2018, and then stay below $3 for the foreseeable future (on an annual basis). The curve will almost always be wrong because weather plays such a big role in gas prices, and there are plenty of traders positioned for a spike this year (i.e. they expect prices to spike much higher that the curve if we have a real winter). But the market is saying the new normal (beyond 2017/18) is sub-$3 nat gas, which is the basis for pricing equities. I tend to broadly agree unless oil prices cause service costs to spike higher which will also affect nat gas prices. In any case, equity prices for non-distressed companies are based on mid to long-term expectation for pricing, so the analysts are in fact using the curve to lower their price targets for nat gas focused E&Ps.

    So do you have a mid-term (3-5 years out) view on nat gas prices that differ from the market?


  4. HS,

    I agree with your analysis of the gas forward curves but I think the mid-term view they represent will be shown to be wrong.

    The Marcellus will peak in the next 3-5 years and that will be a serious wake-up call for those who think that shale gas and tight oil are exempt from the laws of depletion physics. There is nowhere to run from impending loss of 20%+ of total U.S. supply.

    The Utica is a relatively small resource and Antero showed negative PV for Utica PUDs in their 2016 10-K. I will post our recent study of the Haynesville Shale soon but the commercial area is fairly small and already 50% developed.

    That leaves tight oil associated gas to fill the gap. I suspect that the Permian will follow the path of the Bakken and peak because of loss of gas-expansion drive and water production much sooner than most people imagine. Add increased exports and the elements for the next natural gas supply crisis is at hand.

    I may be overly skeptical but the market will adjust quickly to newly found supply insecurity if any part of my story is correct.

    All the best,


  5. Good analysis Art! I have been telling our EPG members for months that I see much tighter natural gas and NGL markets in the U.S. by the end of the 3rd quarter. I believe that the warmer than normal winter of 2015/2016 masked the rapidly growing U.S. gas market. After a closer to normal winter 2016/2017 we have gas in storage closer to the 5-year average today and probably heading to below the 5-year average by October. Increasing exports of LNG by Cheniere and via pipeline to Mexico will come on top of the big increase in demand for power generation coming in Q3.
    In 2018 the gas market may be over-supplied again, but for the coming winter the “gassers” should get good prices.

  6. If conventional gas production is declining by 3.2 bcf/day annually, then the production will fall to zero by 2025. After that point the US would be entirely dependent on shale gas and associated gas from shale oil production for domestic production of gas. If shale gas peaks as well within that time and the decline rate is high in shale gas production…

    Well… there is still some time to figure out a solution.

    Thx for an informative article about nat gas.

  7. Great article as always. Can I ask, how do you calculate comparative inventory? Thank you, M

  8. Maurilio,

    Comparative inventory is calculated by subtracting the 5-year average from the current stock level for the same week. I use a moving average for the 5-year average to normalize the weekly variances.

    All the best,


  9. Art

    I always look forward to your posts
    On your question What Could Go Wrong?

    I have looked at the US EIA data on uses for US gas
    From my interpretation I see the EIA only has four main areas

    When you examine inside all these I am perplexed to see that they are not considering

    Petrochem expansion by my guess at least 3BCF/d based on Methanol/ Ammonia plants sanctioned / relocated from Chile
    LNG which has 12BCF/d sanctioned capacity by 2020 and at least 2-3 BCF/d by end of 2017
    Pipeline gas to Mexico of at least 5BCF/d by 2020 and 3 – 4 BCF/d by end of 2017

    My concern is that if you add these onto the current base use of circa 68 BCF/d US gas supply has to increase to 76 BCF/d end of 2017 and more than 85 BCF/d by 2020

    I realise that there is potentially some switching available from Natural Gas but I think the new power plants sanctioned by my research 2 -3 BCF/d will offset any declines in other areas

    So my point is to What could go wrong is has the EIA taken its eye off the ball and convinced everyone that there forward curves only showing a use of 70 – 75 BCF/d by 2020 has lulled the US market into a false sense of security

    My warning is look at the East Coast Australia gas market – before the LNG plants plenty of low cost gas – today very little very high cost gas

    Your views will be greatly appreciated


  10. John,

    Many thanks for your valuable insights. I agree with you that there are serious potential supply risks when demand is fully understood.

    I have long questioned LNG export from a supply security perspective. I understand that a free-market allows everyone an equal opportunity to do stupid things if they can find someone else’s money to lose doing it.

    I urge readers to learn about the current electric power crisis in Australia that exists in large part because the country allowed unrestricted export of natural gas and now does not have sufficient supply to manage its own power needs. I am not saying that will happen here but think about it.

    There is a magical belief that shale gas and tight oil are exempt from earth physics and the well-established observation of depletion.

    The Marcellus will peak in the next several years and accounts for almost 25% of U.S. gas supply. The Utica is far too small to replace it and Antero–a leading operator in the gas-prone southern core–assigned negative PV to its PUDs in the Utica in its 2016 10-K.

    The Barnett and Fayetteville are in terminal decline.

    The “monster” wells in the Haynesville cost $13 mm to D&C and the $4 commercial core is small and more than 50% developed (a post is forthcoming on those details).

    So, is the future gas (and oil) supply of the United States dependent on the Permian basin?

    All the best,


  11. A offsetting factor has been renewal (excluding hydroelectric) energy additions added throughout last year that resulted in a power burn that is down more than 3 Bcf/d year over year. This is structural with respect to green energy initiatives – and this demand is gone.

  12. TDP,

    I don’t believe it is that simple. I filter out all but wind, solar and geothermal from the “renewables excluding solar” because there is a lot of garbage included like municipal solid waste, landfill gas, and wood.

    What I see for Feb 2017 YOY is the 3.15 bcf/d overall drop that you mentioned in coal, nuclear and natural gas but only a 0.2 bcf/d increase in renewables.


    I think we are seeing overall lower power demand because of a mild winter.

    All the best,

  13. Thanks for the explanation.


  14. Thanks Art- great analysis. Do you see the proposed China trade agreement that includes LNG exports as being a potentially bullish factor in gas prices? Seems if and when enacted, it could reduced the over supply being forecasted. NG prices initially spiked up on the news but have come back down since the report came out early this week.

  15. Jeff,

    I think the LNG trade is based on bogus economics particularly for Asian export. It only makes sense if you believe that landed prices will be much higher in the future or that buyers are willing to pay a premium for supply security.


    No one is paying attention to the re-negotiated gas deal between Russia and China that is underway to bring cheaper pipeline gas to East Asia. Iran has the largest proven gas reserves in the world and will provide gas to South Asia before long.

    All the best,


  16. Art,

    According to the latest EIA drilling report, legacy decline rates are rising faster than production from new wells. Utica follows closely. This is why net production stagnates in Marcellus and Utica and overall US natgas production is still down year over year.

    For that reason the 5.6 bcf/d production growth forecast for the US by end of this year is in my view unrealistic. As Marcellus and Utica mature, legacy decline rates are accelerating exponentially, leaving the net production growth in a deep deficit – despite frantic drilling of new wells.

  17. Heinrich,

    Depletion is inexorable and applies equally to conventional and unconventional reservoirs. Production grows as long as enough new wells are added to overcome the increasing production base decline. Once optimum infill is achieved, terminal decline begins. Rate may be accelerated by over-drilling beyond optimum infill but then, terminal decline will correspondingly accelerate.

    Many people think that shale plays are magically exempt from this fundamental law of physics but they are wrong. The Marcellus will peak and begin to decline in a few years. It represents more than 25% of U.S. dry gas production. There is nothing that can replace it unless some new play is found.

    All the shale gas growth is presently in 3 plays: Marcellus, Utica and Permian. Growth rates are falling in the Marcellus and Utica and may be rolling over in the Permian (by the way, you cannot use the DPR for the Permian because it includes all gas not just shale gas–you must use the Natural Gas Weekly Update that shows the production is about 50% of DPR).

    The EIA STEO forecast for +5.6 bcf/d increase in dry gas production in 2017 seems optimistic but may occur. The real question is, When will depletion take over?

    I suspect it will occur much sooner than most imagine and then we will really begin to question the wisdom of natural gas export.

    All the best,


  18. Hi Art,

    I have followed your thesis on natural gas over the past year (to my benefit as a trader). Gas drive expansion in Permian (loss of reservoir energy) may be a good story for another day be currently the associated gas production from a wide swath of Permian is quite large. Any thoughts on how it might affect the NG supply narrative over the mid-term?

  19. Ryan,

    Associated gas production from Permian tight oil is about 4 bcf/d and will increase as oil output grows. For now, a lot of gas is flared for lack of take-away capacity. In the medium term, much of it may go to Mexico as most Eagle Ford gas does today. I don’t want to minimize the significance of Permian gas but I see it as potentially balancing future declines in the Marcellus assuming that meaningful volumes eventually go into the domestic market.

    All the best,


  20. Thanks Art! Believe it will be something to keep an eye on.


  21. shale gas wells should last over 20 years before terminal death production of 100MCF (100 kilo CF) per day arrives.
    A simple calculation, 15million tons of LNG a year from Gorgon is equivalent to 20Billion Cubic meters of natural gas a year. Assuming production over 30 years accumulates 600Billion Cubic meters of produced gas. With 54Billion dollar initial investment, we have a cost of 0.087 dollar per cubic meter.

    Of course, this huge initial investment include the LNG trains built in the extraterrain in west Australia with super inflated cost. Much like the LNG trains here just started by Cheniere, whose Sabine pass 6 trains costs $10billion for 3.5BCFPD LNG capability, which is about 40Billion Cubic meters of natural gas capability, twice the amount in Gorgon. For 30 years of operation this gives a initial capital cost of 0.006 dollar per cubic meter for domestic shale gas to be converted into LNG.

    The reasonably conservative EUR of a Marcellus well is estimated 6BCF (estimated from the plot generated from the great website by Enno,, or 0.2Billion Cubic meters. With initial investment ranging from 6million to 9million dollars, that’s an initial cost of 0.03~0.045 dollar per cubic meter. Adding the above 0.006 dollar per cubic meter of LNG conversion cost, the American shale gas seems to have an edge over Gorgon.

    In fact, the cost of shale gas in the first major shale gas field in China with annual output over 10Billion cubic meters over 30 years is estimated to cost only 40billion RMB (with 5billion initial exploration cost). That’s 300 Billion cubic meters of natural gas with initial cost of 40Billion RMB or 5.5Billion USD, or 0.018 dolars per cubic meters in the pipeline.

  22. Nuassembly,

    Your point is unclear to me as is the context of your comments? Are you referring to my comment to Jeff and the accompanying chart showing LNG economics?

    It seems that you are making an argument that shale gas is a cheaper supply source for LNG than Gorgon—no argument there.

    That, however, does not mean that LNG makes any economic sense. Your “simple calculation” does not involve any time-value of money for wells or plant construction, nor does it include any costs beyond those items. In any case, $.06 – $0.18/cu meter for plant costs are consistent with the $2-$3/mcf liquefaction costs included in my chart.


    All the best,


  23. Thanks for your elaborations in the reply, Art.

    My calculation shows that shale gas around the world could effectively solve the natural gas shortage problem in places that have the resources, e.g. US and China. This confirms your argument with Jeff that LNG does not have much competitive edge, particularly for projects like Gorgon and Wheatstone where the cost of initial capital has been overblown too much.

    I have found that Cheniere’s 2016 PPT saying that Gorgon project LNG train cost is the highest among all worldwide LNG project at over $2,000/ton capacity (or divided by 30 years, it is over $67/ton or over $1.35/MCF). The Sabine Pass plus CCL 7 trains 31.5mpta capacity will cost $3.4billion, and this put the initial capital cost for 30 year operation at $0.771/MCF. Obviously, my calculations for both shale gas production and LNG liquidification only included initial capital cost and does not include operating cost. This operating cost could make it even fractionated for the LNG liquidification projects, i.e. remote west Australia or Sabine, LA, US. Cheniere’s estimated yearly operating cost is $270million for 31.5 million ton per year LNG Sabine and CCL combined (1.47TCF, or 44.1BCM of natural gas per year), or $0.184/MCF. Assuming the initial capital cost with interest factored in at $1.55/MCF, and the cost or break-even fee will be less than $1.75/MCF — this is lower than the $2-3 range you gave but close. For Cheniere’s $3 fee, they should have a profit margin of $1.25/MCF. For Gorgon, the breakeven cost will probably be higher into the >$3/MCF range. This cost structure will give a good sustainable natural gas price at under $6 FOB Sabine Pass, assuming the Henry Hub price of $3 will be sustainable. This might explain why Cheniere just delivered several LNG cargoes to East Asia, even when the landed prices there reached $5.5.

    I also want to stress that shale gas have cost advantage not only to offshore natural gas but also it comes with capital flexibility. The $3 /MCF upstream cost in your graph probably is now defined by shale gas Henry Hub Price — my calculation shows that most shale gas wells in Marcellus should be able to survive with $3/MCF Henry hub prices even with operating and interest cost. Not only the cost is competitive, but also the resources are probably about 2-3 times higher than you estimated. The plausible average EUR for each well is at least 6BCF. Also, the Utica actually has more resources than Marcellus and upper Devonian has not been fully tapped. With the infrastructure already there, the cost to drill for Utica and upper Devonian could go down quite a bit. I am looking at producible gas at over 180TCF over 30 years with current rate at around 6TCF/year. The top major 5 shale drillers there in Marcellus claim they have core PUD inventory in excess of 30-40 years on average at the 2016 rate and over 20 years on average at the 2014 drilling rate. The other day, I was in West Viginia and a landlord owner told me there is still an unknown Blackriver running underneath Utica.

    I would like to check the current detailed numbers that you believe for Marcellus (with Utica included) 6k wells already drilled these days.
    1. Average EUR , 6~9BCF
    2. Average cost of D&C, $6-9million
    3. Average LOE , $0.5~0.8/MCF
    4. Well head breakeven cost, <$2

  24. Nuassembly,

    I think your LNG break-even prices are unrealistically optimistic. I suggest that you read Michelle Foss’s Is U.S. LNG Competitive?.

    I’m equally unclear about how you derive your Marcellus economics–I can get to $3.50 break-even for core wells but not even close to your sub-$2 numbers. I use $1.40/mcf variable OPEX. I assume you are also excluding royalties and income tax that should be included. I use an average D&C cost of $7.5 mm. Are you discounting your economics? I use an 8% discount rate.

    All the best,


  25. Hello Art, your on going analysis of the Oil and Nat gas market is I the BEST in the world. Art, I think North America is in fro a Big surprise soon because with the increasing demand for Natural Gas and now more LNG exporting NAT Gas together with some of our producers in a decline we will have a perfect storm for a BIG rise in Natural Gas prices

  26. Mundiregina,

    Thanks for your comments. I suspect that the mainstream view of natural gas over-supply and lower prices in 2018 may be accurate. Beyond that, I see little potential for surplus for the reasons that you mentioned.

    All the best,


  27. Art,

    I always enjoy your data driven perspective. Any chance you’ve seen the new WSJ story on how the Delaware Basin drilling is going to create so much associated natural gas that the natural gas market will tank? I’m still not convinced that drilling in the Permian is as profitable as the oil companies claim, especially if services companies can ever regain any pricing power or have to buy new equipment. If the Delaware basin is really profitable in more than just the sweet spots at $45 oil, I guess there will be a tremendous surge in activity and associated gas for a while, but call me a skeptic. Does you data support the WSJ article view?

  28. Was trying to ascertain on Figure 5 how you arrived at surplus/deficits? The EIA balancing item in STEO report does not adequately address it. One could always consider that in a static sense, a deficit is incurred when other sources other than dry gas production from indigenous production, followed by indigenous storage, are relied upon to meet the demand. Net pipeline imports and net LNG imports then fill the void as they are not indigenous.