April 11, 2018 | Kevin Adler & Callie Kolbe
For the last two weeks, Lower 48 U.S. natural gas storage has been below 1,400 billion cubic feet (Bcf), the first time it’s reached that level since May 2014.
The cause of the decline can be summed up in one word: demand. Thanks to a lengthy (though not unusually cold) winter across the Midwest and Northeast, along with record levels of exports, demand in winter 2017/18 has been well above the level of the prior two years.
But unlike winter 2013/2014, when strong demand and falling inventories sent prices surging, the market has basically shrugged its shoulders this year. Prices spiked for a couple of weeks in late-December/early-January, but they quickly returned to their prior range, and have stayed there.
Something is different than it was in 2014, and that’s what we will explore in this issue of Get the Point.
Let’s begin with prices. As the graph below shows, prices began winter 2013/14 (Nov. 1 – April 30) higher than they did in 2018, but not by a huge margin. The Henry Hub price, as recorded by the U.S. Energy Information Administration (EIA), was $3.55/MMBtu (million British thermal units) in the first week of November 2013, and it was $2.78/MMBtu in the first week of November 2017.
Prices in winter 2013/14 rose steadily from $3.35/MMBtu to about $4.39/MMBtu in mid-January and then spiked over the next three weeks to as much as $6.55/MMBtu. By the end of the winter season on the last week of March 2014, they were $4.46/MMBtu, or 25% higher than at the start of winter.
In contrast, prices this winter were flat for the first six weeks of the season and then had a short-lived spike in reaction to record draws from inventories, but quickly calmed down. Prices at the end of March 2018 were at basically the same level as they were at the start of winter—despite the relatively low level of gas in storage.
Traders and analysts have remarked on the low volatility of gas prices in the last few months. In any other year, inventory levels sitting at nearly a 350 Bcf deficit to the prior winter, coupled with a winter season that featured a demand increase of nearly 8.0 Bcf/d from the prior year, would have sent prices surging. It didn’t happen this year.
The graph below shows how low price volatility has been this winter. It plots the storage surplus (positive numbers) or deficit (negative numbers) on the horizontal axis against spot market prices at Henry Hub to get a rough estimate of how the marginal value of the next stored unit of natural gas is correlated to storage inventory. The graph shows how prices in January through March of each year since 2014 have reacted to changes in gas inventory levels. For January-March 2018 (red circles), prices showed little reaction when inventories dipped to nearly 500 Bcf below the rolling five-year average. Only the short-lived spike, which reflects an all-time record withdrawal of 359 Bcf of gas for the week ended Jan. 5, 2018, shows significant volatility.
This is very different than January-March 2014 (purple diamonds) when prices were highly sensitive to inventory levels. In fact, every year since 2014 has seen fairly low sensitivity to inventories.
One possible reason for the lack of market reaction this winter is how we arrived at the sub-1,400 Bcf inventory levels of the last few weeks, compared to about four years ago. When inventories reached 1,380 Bcf on May 23, 2014, it was not the low point of the season. It actually represented the early days of a market correction—the 8th week of a string of 32 consecutive weeks of inventory increases.
The bigger story at the time was two months prior, when on March 24, 2014 storage reached an unfathomably low 824 Bcf. This was the aftermath of a cold winter 2013/14, the first of two consecutive “polar vortex” winters. The market needed a recovery – and it got it, as the inventory rebound through the summer injection season topped out at 3,611 Bcf in November 2014. High prices played a major role by incentivizing production and trimming demand.
Ironically, winter 2017/18 started from almost the exact same point: 3,816 Bcf of storage in the first week of November 2014, and 3,790 Bcf in the first week of November 2018. The graph below shows how the two winter seasons followed similar trajectories for a while, as early-winter showed unusually cold temperatures. Through the first week in February in both years, the inventory levels were within 13 Bcf of each other: 2,184 Bcf in 2014, and 2,197 Bcf in 2018. It could even be argued that with the much higher demand levels in 2018, inventories were under greater pressure this year than four years ago.
Then it changed. Since early February 2018, this winter has seen a flatter rate of decline in storage. But in February, March and April 2014, the decline kept its former trajectory, and the gap between the years widened. For buyers in 2014, alarm bells went off as inventories deviated further from the five-year average, and prices surged.
As stated earlier, gas demand in winter 2017/18 was strong. This was due to several long-term, systemic factors, most prominently export demand and the replacement of aging coal-fired power plants with gas-fired power capacity. Thanks to those factors, even in a winter of moderate winter temperatures, gas demand was higher than it was several years ago.
In fact, this year’s high gas demand did not correlate with a cold winter. The average winter temperature this year was 46 oF, according StatWeather; for comparison, the winter 2013/14 average was 42.7 oF. According to data from the American Gas Association, winter 2013/14 recorded 4,255 heating degree days (HDDs), as compared to 3,781 HDDs this winter.
The impact of cold weather can be seen in the graph below, which compares residential-commercial gas demand in the two winters (as well as other supply and demand factors). In winter 2013/14, gas demand in the res-com sector was 38 Bcf/d, compared to 33.5 Bcf/d in the same period this winter.
But as the graph above shows, other changes in the industry in the last four years are having a greater impact than weather. As noted earlier, exports have reached record levels. The Lower 48 U.S. became a net gas exporter in 2017, compared to importing a net 3.6 Bcf/d in 2014. Moreover, exports are gaining speed, as OPIS PointLogic forecasts a net export level of 1.3 Bcf/d in 2018, with Sabine Pass LNG, Cove Point LNG and (possibly) Elba Island LNG contributing increasing volumes, and new connectivity to Mexico enabling increased volumes to flow by pipeline south of the border. (For more on Mexico, see "Get the Point: As U.S.-Mexico Gas Flows Rise, So Does Clarity on Future.")
Another factor reflected in the graph above is the way that gas-fired power has overtaken coal-fired power. In the power sector, gas demand was higher this season, at 24.4 Bcf/d, compared to 20.2 Bcf/d in winter 2013/14.
The graph below dives more deeply into the impact of the coal-to-gas switching in the power sector. By comparing "power burn per degree", we can see how as more gas-fired power capacity came online and low gas prices made gas-fired power highly competitive, the amount of gas used has been increasing, even at the same temperature level. The high-water mark was winter 2015/16 (green line), when gas prices were below $2.00/MMBtu, but the differential between winters 2013/14 (light blue) and 2017/18 (dark blue) is significant.
Source: OPIS PointLogic
Despite strong demand, gas inventories have not come close to the low point in the early winter of 2013/14. Record gas production is the reason. To use just one point of comparison, in March 2014 U.S. gas production was 60.3 Bcf/d, compared to 78.3 Bcf/d in March 2018—a remarkable gain of almost 30% in just four years.
More gas is coming. Gas production already has increased by 3 Bcf/d since January, and OPIS PointLogic is forecasting that production will reach a record level of 80 Bcf/d this year, on average. In 2014, average production was 70 Bcf/d.
Looking back to four years ago, Appalachia is the biggest change on the production side. Production in the Marcellus and Utica has surged since 2014 from 16.3 Bcf/d to an OPIS PointLogic projection of 27.7 Bcf/d this year. As we look just at this year, the Permian and Haynesville also are expected to contribute significant growth in gas production, moving the market to surplus despite strong demand.
There’s another production factor to consider: the speed at which gas producers can ramp up production to respond to prices. Shale producers, especially in the Marcellus and Utica, have shown that they can shift into high gear if market conditions make it profitable to do so. That’s exactly what they did last year, when the Henry Hub price improved from its 2015-2016 lows to $2.99/MMBtu in 2017. So far in 2018, production in Appalachia has leveled off, but remains near record levels.
High gas production—and the ability to produce even more—are the biggest differences between 2014 and 2018. Gas buyers know that even with strong demand and low inventories, more and more production is coming. All expectations are that production will continue to surpass export and consumption demand, thus allowing for filling storage fields this summer. Production has truly become king.
Location, location, location
Speaking of Appalachia, location matters because of how the gas gets to the areas of highest demand. Nearly 10 Bcf/d of new gas transmission infrastructure in Appalachia has been installed in the last three years (and more is coming in 2018), which supports not only rising gas production, but enables gas to get to markets efficiently and reliably.
Production has increased most strongly in Appalachia, and demand has increased in the areas that Appalachia’s gas can most easily reach: the Mid-Atlantic, New England, the Midwest and the South. In different areas, the story is slightly different, but the general trend is unmistakable: Appalachian gas can meet foreseeable demand growth for power.
And with gas-fired power becoming more prominent, it’s not surprising that where gas is stored has changed from four years ago. The graph below compares where the approximately 1,400 Bcf of gas was allocated across the country on the week ending May 23, 2014, vs the week ending March 30, 2018, using EIA data.
In considering regional storage, it’s also valuable to look at current inventory levels as compared to the five-year average, as they show where the pressure points are this year. The storage drawdowns this winter have been spread across the U.S. Lower 48, with inventories in nearly every region hovering around a 20% deficit to the five-year average, according to the most recent EIA storage report for the week ended March 30, 2018.
The East storage region, which currently sits at a 19% deficit to the five-year average, ended December 2017 at barely a 5% deficit, However, an onslaught of cold weather throughout the last three months brought periods of tightened supply/demand balances, due to higher demand and some wellhead freeze-offs. During these periods, the East storage region relied on increased interregional flows from Canada, the Gulf Coast and the Rockies to help meet demand. In turn, those flows significantly increased withdrawals from salt and non-salt fields in the South Central storage region.
A cold 2018 has also been a feature in the Midwest, where storage inventories currently sit at a 25% deficit to the five-year average. But an important development since winter 2013/14 has changed the market dynamics in that region. Phase 1 of Rover Pipeline came online in late August 2017, and this allows gas from Appalachia to move to the Midwest, thus helping to suppress regional price spikes. Volumes on Rover have been substantial—more than 1 Bcf/d of gas through some interconnects in recent months. Not only has this settled the markets, but Rover’s availability is a key factor to watch as the region moves towards replenishing gas storage during the upcoming summer injection season.
Strong gas demand this winter beat expectations and resulted in a larger-than-expected drawdown in gas inventories. But, as can be seen by Henry Hub prices that remain virtually unchanged from where they began the winter, the market has adapted very effectively. It's not the same story as four years ago, despite national storage inventory levels again dropping below 1,400 Bcf.
Record gas production is the biggest reason why the market remained steady, with Appalachia leading the way. The storage industry itself has adapted, too, as rapid-response salt storage has become a more important and effective supplier to demand markets.
So as we look ahead to the summer injection season, all signs point to a return to inventory levels at the five-year average and continued low volatility in prices. However, the fundamentals surrounding high-turnover salt storage makes it the key EIA storage sub-region to watch this summer for signals of imbalances. Conditions there may likely influence storage behavior elsewhere and, in turn, the direction of prices.
Stay tuned to future editions of Get the Point, for more analysis from OPIS PointLogic of these key trends.