August 15, 2018 | Sadie Fulton
The summer 2018 injection season is more than halfway over, and inventories continue to lag behind the five-year average despite Lower 48 production coming in 9.1 Bcf/d above the prior five-year average. Demand has surged 7.2 Bcf/d above the five-year-average, but that still leaves a supply-demand net increase in supply of 1.9 Bcf/d year-to-date, compared to the five-year average. And yet, injections since the first week of April have averaged only 58% of the prior five-year average injections. This discrepancy places inventories as of Aug. 3, 2018 at 2,354 Bcf compared to last year’s 3,025 or the prior five-year average 2,926.
The orange line in the graph below shows how inventories this year started low and have failed to keep pace with the prior four years.
In late March, if someone told you that inventories would be where they are now, you may have ventured a guess that the market might incentivize injections with stronger prices in the spot and forward market. However, 2018’s prices look sleepier than ever, both in levels and volatility. The three graphs below show how steady the Henry Hub spot and contract price have been this summer, compared to recent years.
Why isn’t the market “worried” about these trend-departing inventory levels? This edition of Get the Point will explore the interaction between supply, demand and storage inventories, looking for red flags that could signal a market that has become seriously out of balance.
Thus far in the year, the phenomenal growth in production has not led to increased injections. The chart below displays how the Lower 48’s total production translates to storage activity on the weekly basis.
As can be quickly noted, 2018’s average daily production lies well outside of the lines drawn by activity in the prior four years. This year’s storage activity, on the other hand, would look perfectly natural if shifted left in the chart. If anything, oddly enough, 2018’s record-setting production levels could conceivably be construed to be correlated with smaller injections and larger withdrawals. So, though production has changed significantly year-over-year, it isn’t dragging injections along with it.
Demand is also trending stronger year-over-year, though with less drama than production. The Lower 48’s total demand drives a much tighter correlation with weekly storage activity than does production, and the logic that more demand means less injection or more withdrawal is clearly supported.
Though production plays no direct role in the image below, its guiding effect can be deciphered. Notice that for weeks with similar average daily demand, 2018’s storage activity is higher than in other years? This implies that for similar levels of demand, 2018 had more potential to inject or less need to withdrawal. This, of course, relates directly to 2018’s production strength.
So, here we are. Inventories are low amid strong production and demand, but prices are flat. Why?
Maybe we are operating under the impression that it would take wild departures from current expectations about winter 2018/19 supply and demand to get the Lower 48 into “serious” inventory trouble. However, the following analysis suggests the departures may need not be so wild.
OPIS PointLogic’s current expectation for end of October 2018 Lower 48 inventories is 3,381 Bcf. We expect winter 2018/19 demand (including exports) to average 102.8 Bcf/d, with average monthly demand peaking at 115.7 Bcf/d during January. We are expecting Lower 48 total supply (including imports) to average 90.5 Bcf/d, again with the monthly average peaking in January at 90.9 Bcf/d. When we replace those expectations with different scenarios of winter 2018/19 supply and demand, we see the magnitude of shocks away from our current expectations it will require to end up short on gas in winter 2018/19.
The following chart gives implied storage in Bcf/d for various scenarios. The scenarios all start from OPIS PointLogic’s current set of expectations and depart weakly or strongly from there. The implied storage figures are then applied over the whole season to arrive at end-of-March inventories for each of the various scenarios, which are also charted below.
The most plausible scenarios involve demand and supply moving together. These scenarios are highlighted below in bold font. The range of possibilities around end-of-March storage inventories vary widely, depending both on supply and demand, but also on where end-of-October inventory actually falls.
To use the tables below, start with OPIS PointLogic’s current expectation of both supply and demand. In the implied table, the expectation starts at -12.3 Bcf/d. In the end-of-March inventories table, the expectation is 1,524 Bcf. Moving left from there, average daily supply is reduced by 10%, 20% or 50%. Moving right represents supply increases away from our current expectation. Expectations of average daily demand are reduced by moving up, or increased by moving down.
On top of this, OPIS PointLogic is increasingly aware of low-side risk to our current projection of end-of-October inventories. The tables below are a rerun of the first tables, but starting from a baseline of 3,100 Bcf of gas as of Nov. 1, 2018, instead of 3,381 Bcf. While this number is well below any forecast right now, it’s not out of the question if we have a late-summer/early-fall heat wave in high-population areas.
In conclusion, there are several plausible-enough scenarios of winter 2018/19 supply and demand that leave March 31, 2019 inventories in an unprecedentedly low place. If Oct. 31, 2018 inventories come in at OPIS PointLogic’s current expectations, supply comes in as currently expected, and average daily demand sees a 10% increase, inventories by the end of March would be completely depleted. Same goes for a 10% reduction in daily supply all else as expected.
Also, for additional details about the unusual inventory situation -- including a look at regional flows and drilled but uncompleted wells -- read the related IHS Markit/OPIS PointLogic special report, "Safe Bet on Supply?"
This is a winter season to watch – OPIS PointLogic will continue to follow these trends. Stay tuned.