« Back to Get the Point main page

Did Salt Storage Operators Head Fake the Market?

June 20, 2016 | By Luke Larsen

There are more than 400 natural gas storage facilities in the U.S., and they currently contain about 3 trillion cubic feet (tcf) of processed and deliverable natural gas. This is known as the nation’s working gas inventory.

In the past, this 3.0 Tcf of natural gas was considered the benchmark requirement by operators and industry analysts of natural gas on-hand in order for the U.S. to survive each winter heating season. However, this outdated measurement comes from a time when there was very little use for natural gas outside of a declining industrial base and relatively stagnant use among residential and commercial consumers. In this issue of “Get the Point,” we explore how changes in gas demand for power generation has changed the dynamics of gas storage in the U.S.

Over the course of the past two decades, power generation has acquired a larger and larger share of the natural gas market, which has forced the industry to up the ante on annual storage inventory fill requirements. So what is the new benchmark requirement for winter? Is it 3.5, or perhaps 3.7 Tcf? It’s tough to say, but it does seem that for the most part, market forces of late have pushed to the upper echelon of available storage inventory and capacity levels over the past several years. 

Annual Maximum Inventory prior to First Draw

As Figure 1 demonstrates, the increase in storage gas available prior to the first winter drawdown is founded on the fact that the balance of production, demand, imports and exports has pushed inventories to record-setting inventory levels. As producers reveled in an extended period of a high-price commodity environment a few years ago, the internationally priced incentives led them to consistently ramp up capital expenditures (CAPEX) and push supply to levels the U.S. had never seen before.

Start a Free Trial of PointLogic Energy's Pipeline Flow Data
This evolution in supply has all of the demand variables teetering atop a production bubble -- waiting for the proverbial next shoe to drop. And whether it is LNG or Mexican export infrastructure development, a slip in production levels or a change in the economics of the power generation stack, each element waits to make its presence known in the overall U.S. natural gas supply and demand balance.

In the last two weeks, the market has seen a rally in gas futures of more than 20% to bring the July prompt month contract above $2.60 /MMBtu. Back in early March, the July contract (which is typically one of the more premium summer months) was trading just below $2.00/MMBtu.

An interesting side note regarding the shape of the forward curve suggests continued nervousness with length in the front of the strip, with July remaining discounted to October by more than a dime. This relationship is revealing because it shows that spot prices continue to lag somewhat behind prompt futures. The OPIS/PointLogic Henry Hub price index for next day delivery has averaged just shy of $2.35/MMBtu for June and has been disconnected from the prompt July 2016 contract by more than a dime most of the time, but this has converged ever so slightly over the past few days.

The chart below illustrates the physical to futures phenomenon.

Recent Spot to Futures Relationship

A longer-dated trend analysis of this relationship reflects that for the most part Henry Hub and prompt gas do move in tandem. OPIS/PointLogic data supports the chart below for all of 2016. We must note that the ‘June’ referenced gas futures prices in the table of Figure 2 above does not actually represent a daily settlement. These are PointLogic proprietary observances of the postings as cash trading comes to a close each day.

The graph below represents the true daily settlements of the prompt (current) month over the course of 2016. What is very evident is the late-December freeze off induced rally from the $1.75/MMBtu low, as well as the latest rally in the July contract from the below-$2.00/MMBtu threshold in late May. Overall, this area has offered firm support since mid-April.

Henry Hub Natural Gas Front Month Futures

Now, we take a look at the OPIS/PointLogic Henry Hub spot price index for the same timeframe.

OPIS Henry Hub Index

Figure 4 represents physical spot transactions at Sabine’s Henry Hub, and for the most part emulates the price movement in the futures contract -- but it clearly stays somewhat disconnected from the prompt futures. The exception from the relationship was in early January, when the spot price nearly converged to the February contract at the time.

However, comparing the prices in late May with the July contract shows that the contango, or discount, relationship diverged further over the extended Memorial Day trading period. Throwing out the two noticeably large price rallies that were preceded by low-demand holiday trading periods (subsequently inspiring numerous storage operation restrictions from an injection standpoint), the fundamental element that generally keeps cash prices closely tethered to the near term futures are salt storage facilities.

Salt Storage Facilities

Understand the relationship of spot price to futures price must include a look at salt storage situation. The Energy Administration Information (EIA) reported a salt storage balance of 370 billion cubic feet (Bcf) as of June 3, 2016. 

Salt storage is important because it is the gas that can most quickly be both supplied and refilled – that is, it’s the most responsive to spot market forces. Speedy product movement creates a significant influence, even though it represents less than 15% of total storage capacity in the nation.

The estimated capacity of salt storage is around 470 Bcf. While we appear to have not yet bumped up against any fundamental capacity issues with these types of caverns as of yet, salt storage is about 86% full at this time, and PointLogic notes that we are only 11 weeks into a 32-week “Summer” injection season. It would seem logical that the actual number of weeks that we recognize working gas inventory gains generally ends up taking more shoulder weeks, due to the current oversupply situation in the market. However, reviewing EIA’s weekly history over the course of the modern day shale period (calendar years 2008 through 2015), there is no noticeable direct correlation between the huge growth in production and the number of storage injection weeks on an annualized basis.

Since 2008 the U.S. injects (or builds) working gas levels an average of 34 weeks out of the year and draws down during the other 18. Perhaps the only revealing element of this data range is the modest impact of an additional 2-3 weeks of drawdowns during both the calendar years of 2013 and 2014.

Generally speaking, the number of injection weeks and withdrawal weeks continues to be closely represented by the seasonal make-up of the natural gas industry. ”Summer” spans from April through October (32 weeks), and “Winter” is 20 weeks (November through March).

52 Week Annual Review

When looking at salt storage facilities, the weekly drawdowns-to-injections ratio over the course of the year, we see a steady pattern of 20 Winter and 32 Summer weeks over the last five years, on average. However, because most of the salt facilities being utilized are in the Gulf Coast states and represent less than 15% of total available U.S. storage, the actual weeks when the Winter and Summer occur in the salt storage facilities is different on a regional basis. This is because injections generally occur throughout a larger number of shoulder seasons for weeks at a time because the facilities in the northern portions of the U.S. are in a drawdown mode during both of those periods. On the flip side, it should be noted that these additions are somewhat offset by a period of summer withdrawals, as these storage services are heavily relied upon to serve peak power generation needs because of their flexibility.

This is a key point, so it is worth repeating: deliverability from salt caverns is typically much higher than other storage reservoirs. Inventories can be withdrawn and/or filled in a much shorter period of time.  Peak power generators desire these services is because operators can begin flowing gas in as little as one hour’s notice.

2011-2015 Average Weekly Storage Deltas

This reliance upon salt storage flexibility results in a short duration of drawdowns during the peak summer weeks. So for a five-week period from mid-July to mid-August, we have typically seen a reduction in salt storage levels. From a salt storage inventory perspective in 2016, this yearly drawdown can’t come soon enough.

The current salt working gas inventory level of 370 Bcf reflects a surplus of 140 Bcf, vs. the five-year average and is also ahead of last year’s inventory by 83 Bcf. Annual salt inventory levels have also been living in the upper echelon of capacity levels, with a peak of 381 Bcf achieved in November2015. And while in more normal winters we typically would see 15 weeks of drawdowns totaling more than 140 Bcf of working gas reductions, this past winter had only 13 weeks of drawdowns totaling 101 Bcf of gas. This was followed by 13 weeks of injections that began the last week of February that added back 90 Bcf of gas at a time when, historically, we should have only seen working salt inventories increase by 56 Bcf.

This recently established maximum storage mark would have been exceeded some time in the past two report weeks if average injections would have held true to form at the average increase of 14 Bcf over the last five years. However, salt operators instead chose to draw down inventories the week ending May 27, as EIA reported a decline of 2 Bcf. Furthermore, the widest cash price relationship to a forward month that we had seen in some time showed net inventories for that Memorial Holiday week coming in unchanged, which defies all logic associated with the management of expensive salt services. Generally, marketing or asset teams will seek to turn over as many of these salt-stored molecules as possible in order to cover their higher costs.

Current PointLogic analysis shows that we will likely see a larger week-on-week injection overall in the South Central region, with a +15 Bcf increase, or +4 Bcf compared to the prior week. This increase is supported by the fact that the transparent facility injections actually doubled for the week ending June 10. 

The strange phenomenon that resulted in the 2 Bcf drawdown for the week ending May 27 is also identified through PointLogic flow data, and this market-shaping movement was perhaps heavily supported by the speculative nature of operators who seized the opportunity to create space for a better opportunity from a baseload standpoint in June. This gamble will only work if demand remains heavy for the entire month of June and continues into July.

For now the question remains: Did salt storage operators hoodwink the market near the end of May based on operational needs or out of speculative envy?

Stay tuned to PointLogic as we continue to monitor salt storage activity and its unique ability to help balance the supply and demand forces shaping price over the summer.  

« Back to Get the Point main page


Sign up here to have Get the Point delivered to you each week!





Share With Your Colleagues






© PointLogic Energy, part of OPIS an IHS Company | Site Map