What Does An ‘OIL GLUT’ really mean?

by Brynne Kelly

July 24, 2017

What Does An ‘OIL GLUT’ really mean?

The terms “glut” and “oil” have become almost inseparable the past 18 months.  Fueled by an obsessive focus on inventory and production levels, it has become an easy narrative.

The Cambridge English Dictionary defines a ‘glut’ as:

A supply of something that is much greater than can be sold or is needed or wanted.

A glut is often associated with low prices, or an expectation of low prices.  An extreme example of a supply ‘glut’ can be seen in the electricity markets through real-time electricity prices.  Whereas most commodity markets exist with a storage system under-pinning them,  this is not the case for electricity markets (but for limited battery and hydro storage).  As a result, when power has been generated in excess of demand, market prices have to move low enough to clear this excess supply.  As a matter of fact, this market clearing price can even be less than zero (negative), meaning the producer has to PAY the consumer to take the product off their hands.  The negative price reflects the ‘cost’ the consumer incurred to accommodate that supply into their system.  Without a storage system in place, market regulations require utilities to hold enough generating capacity to cover the highest projected demand for the year plus a reserve margin.

A robust storage system helps to smooth day-to-day and seasonal differences in supply and demand and reduces traditional production capacity requirements.  In this way, storage acts as an alternative form of supply and demand by allowing product injections during times of over-production for use in times of under-production.  The natural gas market in the US is a good example of this.  Total natural gas production capacity exceeds total demand in the summer, but isn’t enough to meet total demand in the winter.  Using storage however, production can continue at a relatively stable level throughout the year, with excess being injected into storage in the summer for use in the winter.

Storage levels then become the ‘glut’ indicator, unlike the electricity market which doesn’t have a buffer between differences in supply and demand.  At the extremes, conventional storage systems either become too full to accommodate additional supply, or too low to accommodate projected future seasonal demand.  In either case, price signals emerge in an attempt to rectify the situation.

Perceived versus Real “Glut”

An inventory shortage or excess can be either perceived or real.  Real inventory issues show up in the spot market.  Perceived inventory issues are based on assumptions.  These assumptions are anticipated as threats to the spot market at some point in the future and are expressed via the futures market.  If these assumptions don’t manifest into the spot market over time, assumptions regarding the ‘future’ will ultimately be revised.  Think of flowers delivered on Valentine’s Day.  A month in advance, the market establishes a clearing price for flowers delivered exactly on February 14th.  There are even discounts offered for flowers delivered the day before or the day after February 14.  These are based on perceived value.  Come Valentine’s day, those who forgot to place their order and need flowers are now forced into the ‘spot’ market and may be willing to pay multiples of the original price for delivered flowers, or they may luck out and find an overzealous store stuck with too much delivery capacity.

With that in mind, let’s circle back to the aforementioned ‘glut’ in the oil markets.  The prevailing narrative is that there is too much supply for the current level of demand.  The evidence being used to support this narrative is inventory levels.  High inventory levels are perceived as a threat to the spot market.  As long as there is still room to fill conventional storage, producers have the option of storing their barrels or selling them.  In this way, spot prices are linked to futures prices.  In a perfect world, the difference (spread) between the spot price and the next futures price should reflect the cost of storage capacity.  Of course, the cheapest storage options will be used first.  As these fill up, spreads may widen to value the next most economic form of storage and so on.  This is why it has been traditionally assumed that a widening spread, or contango, may be signaling a market oversupply.  On the other hand, if spot prices are higher than futures prices, those with barrels in storage will withdraw stored barrels to capture a higher spot price and eliminate storage costs.

Take a look at the 1-month price spreads in WTI futures since the beginning of 2017:

At the beginning of the year, the front of the futures market was trading at a slight contango from one month to the next (the black line above).  However, starting with the Dec-17 contract, the market had moved into ‘backwardation’.  This market structure makes the decision to store barrels uneconomic and should have induced producers to sell their barrels into the spot market.  You can see that by June, the backwardation in the futures curve had been pushed back 3 months (the grey line) and by last week that backwardation was gone.

The effect of the shape of the term structure seems to have had the desired effect on inventories, as storage has been drawn down significantly since April:

Total Petroleum inventories have been reduced by over 50 million barrels since the beginning of April, 2017.  This is significantly higher than inventory withdrawals in the prior 3 years.  As a result, total inventory levels are now slightly below where they were at this time last year but still above levels of the prior 3 years:

Once the incentive to store barrels was reduced, they were instead pulled out of storage and moved to market.  Pulling over 50+ million barrels out and selling them into the market put pressure on the front of the futures curve relative to the back (seen in the chart below):

As a result, the December 2018 contract is now close to $2.00 over the December 2017 contract.  While that may sound like a lot, it’s just a little over $0.15 a month compensation to defer payment on your production.  Since the seller of oil doesn’t receive payment until after the oil is delivered, holding it in inventory doesn’t generate cash.  As a matter of fact, holding oil in inventory costs money (via the tank storage cost).  With all the news coverage this past year regarding the need for cash, it’s no surprise producers would rather sell now versus hold inventory.

Days of Supply

Oil inventories are often expressed as “Days of Supply”.  This is essentially total inventories divided by daily demand.  Over the past 10 years, we have gone from just under 25 days of supply to a little over 30 days as production has increased and demand has flattened (as seen below).

This assumes oil inventories can transported to refineries to be processed as they are withdrawn.  I use ‘days of supply’ to put the phrase ‘inventory glut’ into perspective.  Is just over a month’s worth of crude oil held in inventory to meet our current demand needs something you would consider excessive?  Since we still rely on imports for over half of our crude oil needs in the US, this doesn’t scream ‘excess’ to me.  But, as in any business, holding inventory is expensive and it’s a delicate balance to determine the optimal balance between margin and cost.  Assuming there are decent margins to be made, it would be logical to assume more inventory would be held at low prices versus high prices.

What is The Signal?

As Nate Silver discussed in his book “The Signal and the Noise” (2012), we seek to extract signals and eliminate noise when building models with historical data to
predict the future.  However,  while the final outcome of an event is subject to influences that will repeat themselves in a foreseeable way in the future, it is also subject to a great deal of noise that will not repeat itself in the future in a predictable manner.

I believe the ‘signals’ we have been receiving from the oil market have moved from ‘price’ signals to ‘volatility’ signals.

Increased production and storage capacity in the US have reset oil prices from $90 to $45 in the past 2 years. As a result, production costs have been targeted and optimized.  What remains is a larger, more efficient system with a lower convenience yield that may be less prone to significant price moves in either direction.

However, low volatility also makes it more difficult to justify new assets as it reduces projected returns (all other things being equal).  With long lead-times to build new assets like pipelines and export infrastructure, a sustained period of low volatility will have impacts not realized for many years in the future.





Understanding The “Mexico Deal”

July 5, 2017

Understanding The “Mexico Deal”


So, you’ve heard about this large oil transaction that has the potential to move markets and wanted to know more?  In the article below, I use this transaction as a means to delve in to the structure of the type of trading group that might participate in this deal. From there, I then take a deeper look at the transaction itself and the complex hedging decisions those that participate might face.  With this information, you will learn where to look for any related trading opportunities.

We know that several years ago the Mexican government implemented an annual crude oil hedging program as part of their budget process.  Their intent is to protect the oil revenue defined in their yearly budget through the purchase of put options.

Recently, knowledge of this deal has received more press and it is now widely anticipated as a transaction large enough to move markets.  To understand the process that occurs from deal negotiation through deal closing, let’s first take a look at the structure of the trading desk.

Click here to receive future analysis

Trading desk structure

To understand how large deals like this unfold, it’s helpful to understand the different roles on a large integrated trading desk, specifically the trader and the marketer roles.

You may have noticed that many of the large banks and integrated energy companies out there have what they call their “Trading and Marketing” group.  That’s because they are actually two distinct functions that depend on each other to make money.

Trading and Marketing roles defined

The trading desk manages a portfolio of risk exposure, executing strategies to capture anticipated market moves.  They are usually governed by Value at Risk (VAR), Volume, and Credit limits.  VAR is a common measure of how much a portfolio might lose (gain) given normal market conditions for a given confidence interval (probability) and liquidation period.  It is a measure of market risk.

When working inside a large institution where capital constraints like margin are monitored at the corporate level, it’s more relevant to assign VAR limits at the individual trader or desk-level rather than assign margin limits.

In addition to VAR limits, traders or trading groups are often assigned volume limits.  These volume limits are often assigned the total portfolio level, the individual market level, or both.  The rationale for volume limits is to prevent position size from impacting the liquidity-portion of the VAR calculation, among other things.  It’s not an exact science, and in my career I have certainly run up against assigned volume limits before my VAR limits.

The introduction of central clearing has simplified portfolio credit risk in listed products, however, it is still an issue for non-listed products.  There is still a vast world of over-the-counter (OTC) transactions with non-standard terms, non-standard durations, non-standard grades and variable volume requirements.  These transactions are enabled by direct credit agreements negotiated between counter-parties in the deal.  As a result, corporate or portfolio-level credit exposure limits by counterparty are established to ensure exposure to a particular counterparty doesn’t exceed collateral agreements.

The marketing desk in a trading business usually focuses on originating non-standard transactions directly with another counterparty.  Typically a marketer within an energy trading group might be assigned a region of the US and work to build relationships with both producers and consumers within that region.  For example, they may have an industrial client that is looking to buy 800 Mmbtu’s of natural gas delivered to their plant location via pipeline in southern Alabama.  This is not the standard 10,000 Mmbtu exchange-traded natural gas future.  Since the marketing role is to originate business directly with customers, versus take risk, the marketer will bring the terms of the deal to their internal trading group for them to quote a price.

There can be a lot of discourse between the trading and marketing groups regarding deal valuation.  Marketers generally don’t manage market risk, they manage relationships. They are primarily measured on the amount and quality of the deals they close.  Conversely, a trader is measured on the returns they generate from the portfolio they manage.  Once a deal is closed, it is transferred to the trading desk to manage.  Some groups establish a ‘transfer price’ at which the deal is moved into the trading ‘book’ from the marketing ‘book’ in an attempt to capture any value (a marketer’s negotiation skills) added to the sales/purchase price over and above the trader’s quoted price. Ultimately, the more deals the group as a whole gets a look at the more they are aware of what is going on in the market, which is valuable information.

It would be reasonable to assume that the opportunity to participate in the Mexican hedge deal originated through the marketing group and the relationships they have created over the years.  It’s also reasonable to assume that these marketers work closely with their internal trading desks regarding the terms of the deal so that their traders can establish a market price at which they are willing to transact.

Why do I assume this?  Deals like the Mexican oil hedge require a lot of contracts with specific negotiated terms (credit, margin, legal, etc).  This requires a lot of time, which is something that traders on a desk don’t usually have.  This is clearly a deal done directly between two counter-parties, contains non-standard terms (if it was a simple put option on WTI futures, it could be executed via the exchange) and its contract language would have to be carefully monitored and updated each year.

The benefit of being a Market-Maker

Market-making on a trading and marketing desk, as described above, refers to the internal process of the traders providing market prices to the group’s marketers on deal opportunities they originate from their customers.  Once the marketer has obtained a price from their trading desk and established how long that price is good for (1 hour, 1 day, etc.) they submit them to the customer.  Often times it is unknown how many other companies the customer is soliciting markets from.  Because markets are submitted directly to the customer, they are not public which makes for a distorted feedback loop.  You generally have no idea what prices your competition gave to the customer, but can only assume that if you didn’t win the deal, you did not provide the best price.

Over the years, I have worked with marketers and participated in pricing long-term, non-standard deals.  In most cases the counterparty needs either variable volumes, obscure delivery points or an illiquid grade of product.  The counterparty has often reached out to several companies to ensure they are getting competitive prices.  This is why an active trading desk is needed as part of an overall group.  Being active in the physical and financial markets every day gives traders the knowledge needed to understand market dynamics and the performance risk they are being asked to ‘price’.

One thing is for sure:  the fewer market-makers and the less liquid the product being priced, the more opportunity there is for adding larger premiums to your markets.

It’s fairly common for markets to have large annual or seasonal transactions that occur due to regulatory requirements, hedging or procurement. Some of these transactions are governed by regulatory rules regarding the type of counterparty that can participate.  These rules are typically related to the size and credit-worthiness of the participants.  Limitations on the number of participants limits the amount of competition and increases price mark-ups (mark-downs).

Which brings me back to the annual hedge by the Mexican government, as the above speaks to the rumored ‘large fees’ made by those that qualify to make markets for this deal.  I believe “large fees” relates to how wide one can make their market.

Due to the size of this deal, and the potential pay-outs, Mexico is interested in dealing with counter-parties that are well collateralized with the ability to pay out large sums in the future should their option go ‘in-the-money’. There is nothing worse than having a winning trade on your books only to have your counterparty go bankrupt or become unable to pay (as happened to many when Enron filed for bankruptcy).

Electricity Market Example

To illustrate that deals like this occur across all industries, let’s take an example from the electricity market that I am familiar with.  The electricity market is regional with listed futures contracts for major pricing hubs within each region.

There are two main listed futures for each pricing hub:  On-peak and Off-peak think high-sulfur and low-sulfur crude oil grades).  In the eastern time-zone, On-peak contracts cover the 16 hours (7 AM-10 PM) on each week-day of the month. The Off-peak contract covers the 8 hours (10 PM-7 AM) on each week-day of the month and also the entire 24 hours for each weekend and holiday in a month.  The volume for each of these contracts is, for the most part, 50 MW/hr multiplied by the number of On or Off-peak hours in each month.

An electricity marketer communicates with many different types of customers to originate a deal (large industrial users, builders of new generation, etc.) When a customer is looking to transact, they provide the terms of what they are looking for and the deadline for prices to be submitted.  In this example, let’s say that the customer is a large industrial company that is growing and needs to purchase additional electricity to cover that growth.  Their plant is located in Pennsylvania and they provide the following estimated usage curve indicating how much they are looking to buy each hour:

The customer has requested a fixed price quote (given this usage curve) for a 5 year term.  The volume of the futures contracts traded in the market are for 50 MW/hr blocks, while the volume the customer needs per hour varies throughout the day. Hedging this deal using futures contracts will leave the trader with excess during some hours and shortages in others.

Just as demand varies by hour, so do prices (as illustrated in the graph below).

You can see that using the standardized futures contract as a hedge would leave the trading book with excess length during the hours of the day that prices tend to be the lowest, and with shortages with during the hours of the day that prices tend to be the highest.

The trader will factor this into their model when crafting their offer price.

Here are a few basic hedging scenarios that might take place if they win the deal and it ends up in the trading book (as a sale to the industrial customer):

  • Buy a 50 MW block of on-peak futures (7 AM-10 PM, Mon-Fri) and stay short the off-peak hours (nights and weekends),
  • Buy a 50 MW block of both On and Off-peak futures which makes the net position fairly flat during on-peak hours, but net long during off-peak hours,
  • Do nothing – it’s a nice compliment to the trader’s overall bearish positioning and view that futures prices are headed lower

Of course, there are other options besides using futures contracts.  This deal may be somewhat offsetting to an existing position already in the trading book (in which case the trader might have an incentive to offer a better price to the customer because it relieves some headaches and locks in some profit).  The trader could also utilize a combination of options, however, options on electricity futures that settle against the hourly price averages are fairly expensive due to the higher volatility of hourly prices versus daily.

Other factors that affect hedge activity relate to corporate-imposed risk limits (volume and VAR limits).  Some deals are large enough that they would cause the trading portfolio to exceed their risk limits, if not hedged immediately.  Of course, you can see from the example deal above that it’s not always possible to hedge ALL of the risks immediately in some deals (like the variable hourly volumes and prices).  However, if the size of the new deal is enough to put the portfolio over volume or VAR limits,action must be taken immediately using liquid futures.  Post-mortem analysis will determine what net risk is left in the portfolio to manage (i.e. basis, location, quality, volume, etc.).

This is when a trader assesses their ‘net risk’.  Ending up long basis (spread) as a result of hedging is still a trade.  Would you buy that basis (go long) regardless?  If not, sell basis.

Deconstructing the Mexican deal:

Using the knowledge of the overall dynamics reviewed above, let’s look at how that might apply to the Mexican deal.

We know that roughly 95% of the crude oil that Mexico sells to the US is Maya heavy crude for use by Gulf Coast refiners.  Pemex (Mexico’s state-owned petroleum company) calculates their sales as a derivative of other market prices.   Their current formulas are:

Since over 95% of Mexico’s crude oil sales to the US are of Maya crude, that’s the formula we will look at for this analysis.  Maya’s value in the Gulf Coast is  indexed to a basket of 3 crude oil grades and 1 fuel oil grade, plus a constant (K) commonly called the K-factor.  This constant is updated monthly and posted on their website (found here) as seen below:

There are a lot of moving parts, each having different sets of price influences and drivers.  Also interesting to note,  WTI is not used directly in any of these formulas.  This is what Pemex states on their website regarding their pricing formulas:

First, let’s use what we have learned so far to evaluate the deal that was done for 2017 (the deal runs annually from Dec 1 – Nov 30).

Deal Assumptions: 

(see August 29, 2016 WSJ article)




The term of the deal has historically been 12 months running from December to the following November.  Using all of this information, it’s fairly straight-forward to pull together a back of the envelope analysis of the 2017 performance of $38.00 options against the market:

I pulled monthly posted prices (for US delivered Maya crude) from the EIA website for the beginning of 2017 and used the Pemex formula for the balance of the year.  It doesn’t appear that any of these puts would be in the money.  Given that market prices are only modestly above the strike, this is not great news for Mexico since they are net long oil. The premium they paid would be much easier to stomach if oil prices had moved much higher.  Cashing in on their puts is not actually the goal here since this is basically catastrophe insurance.  Any ‘gain’ derived from this insurance really just represents an overall loss of revenue on their total oil production.

With that in mind, take a look at Maya oil prices in the chart below:

With the Pemex formula and the futures curves for each of the 4 benchmarks, you can get a feel for 2018 Maya ‘futures’.  What strike level, if any, will they consider for next year?

For reference, futures curves can be seen in the table below:

For those who are more ‘visual’, here are the futures curves as of July 5, 2017:

The unknown in the calculations above for future terms (besides movements in outright price) is the Pemex published “K-factor”.  I used the posted K-factor for June 2017 (shown earlier) as a constant for the balance of the year and 2018.  While Pemex is responsible for setting the K-factor, it seems to be loosely correlated to the LLS/Brent spread.

Analyzing the Risks to the Market Participants in the Annual Mexican Hedge Deal

A recent Bloomberg article shed a lot of light onto this transaction.  One thing the Bloomberg author noted was that changes in bank regulations may be impacting post-deal activities:

Putting aside that a bank may have multiple trading groups and portfolios that ‘take the other side’ of any hedges they transact, I will focus on the market risk that a single portfolio involved in this deal might face rather than the corporate-level net portfolio.

We know that in the past the options purchased by the Mexican Government are primarily based on the underlying price of Maya, and to a lesser-extent, Brent.  I will assume that would again be the case for 2018.

As already mentioned, we know that Maya’s price is derived as a percentage of WTS, High sulfur fuel oil (HSFO), LLS and Dated Brent prices plus the variable constant set by Pemex.  This gives us the ability to understand the risks the seller of these options will need to lay-off once the deal is done.

Primary price exposure:

  • WTS
  • LLS
  • Dated Brent

Secondary price exposure created as a result of hedging (since it’s highly unlikely the seller can equally offset their exposure using Maya futures):

  • WTI/Brent Spread
  • LLS/Brent Spread
  • Brent/Oman Spread
  • Dated Brent/Brent Futures
  • Light/Heavy Crude Oil Spread
  • WTS/Canadian Heavy Spread
  • Etc.

For example, if risk were entirely laid-off using WTI futures the book is now exposed to any significant change in WTI’s relationship to Brent.  Another risk the seller of the options might incur relates to the expiration differences between monthly and average price options (it’s been noted that the deal is comprised of a basket of Asian, or average-price options).  To the extent that European or American options are purchased as a hedge, the option seller will still have to manage the difference between hedges using monthly strikes and the average of daily Index postings that are used to settle Asian options.

Looking at the market risks identified above, you get a sense of how complicated a hedge strategy could become.  Any unexpected shift-change in the relationship of the hedge contract to the underlying products used to price Maya (i.e. WTI suddenly trading at a premium to Brent) could seriously impact the hedge effectiveness.

It’s unlikely that there is a deep, liquid market for over-the-counter Asian options on Mayan Crude.  Therefore, hedging this deal will require the use of a mix of more liquid futures and options markets.  Determining the optimal mix of products to use is an art and a science.  The ‘art’ being any market bias regarding price direction and spreads.  The ‘science’ being the use of statistical tools and models.

How would you decide to hedge this trade?  Two statistical tools that are often used when evaluating effective hedge markets are ‘correlation’ and ‘r-square’.  Correlation measures the strength and the direction of a linear relationship between variables.  R-square measures the proportion of the fluctuation of one variable that is predicted from another variable(s).

Shown below are two simple correlation tables for the markets used in the Maya pricing formula (based on 2 different sets of historical price data):


In both time series, one thing that stands out is the high correlation between Maya and LLS oil price moves.

Since WTI and Brent are the two most liquid futures contracts, I went on to include historical WTI prices in the mix (even though WTI isn’t specifically included in the Maya price formula) and ran correlations again, using the same two historical time periods:

With the addition of WTI, we reveal a high correlation between WTI and WTS (which is a main component in the Maya price formula).




Price correlations provide useful insight.  To get more specific however, the R-square coefficient is used to define the usefulness of those correlations.  R-square is a measure of how much the variance of ‘y’, or in our case “Maya”, is explained by the model of continuous predictors “x” (in our case WTI, Brent, LLS, WTS, HSFO)

R-square outputs (using the same historical time series as presented in the correlation matrix) are shown below:

Each bar represents the historical price series used as well as which “index” variable is being compared to Maya.  For example, the first blue bar on the far left is the R-square of Maya and LLS prices using historical prices from 2015 through 2017.

Notice the difference in the results of the two historical time series used.  Specifically, the decline in the r-square of Maya vs HSFO (labeled R^2 HSFO above) in recent years.

This is obviously one of the problems with using historical time series data.  The changing nature of spread relationships can be significant and render older data useless.  Two years ago, heavy crude oil was pricing significantly below lighter grades, however, those relationships have changed with the OPEC cuts that started last year (which took the production of heavier grades off the market raising their price relative to lighter grades).

We could go on and on using various statistical models, but you get the idea.  The point is to get a sense of how those involved in the deal may look to hedge.  With both LLS and Brent showing the highest correlation and therefore, r-square values with respect to Maya, WTI may not be the hedge of choice.  With WTI as one of the more depressed light grades in the market lately, using it as a hedge leaves the portfolio essentially “long” Brent and LLS (the result of selling puts on Maya crude) and short WTI.

Said another way, the trader who has hedged with WTI has now made a market call on the Brent/WTI spread.  Selling put options (i.e. to the Mexican government) creates ‘length’ in a portfolio.  Think of this length as being related to Brent and LLS (and also WTS and HSFO).  Selling WTI futures to reduce

this length leaves the portfolio with a spread, namely long the Brent/WTI spread or long the LLS/WTI spread.

The knowledge of how such a large deal may be hedged can lead to trading opportunities.  For example, if I believed the initial hedges might be placed in the most liquid markets such as WTI, I might expect the Brent/WTI spread to widen in response to heavier-than-normal WTI futures selling.  Therefore, it might pay to go long that spread before hedging begins. The trader responsible for hedging this deal might also do the same hoping to pull some more profit out of the market from trading positions.  You can extrapolate this concept to High vs Low sulfur fuel oil, Gulf coast vs New York harbor spreads, etc.  My point is that outright price movement isn’t the only outcome here, and you might be well-served to look at how the various spread relationships are behaving.

Since the initial move in spreads after the OPEC cuts last year, the entire oil complex has remained closely linked.  A break-out move in any of these spread relationships might be a signal that significant hedging volume has entered the market.  Pay close attention.  Anyone who has significant volume to sell in the market might hit those that are less liquid first to ensure they can get some sales off before the market senses what’s coming.  Since selling pressure in WTI and Brent can pull the entire complex down with it, one trick traders use is to sell or go short some off the less obvious markets (WTS, HSFO, etc.) beforehand in hopes of capturing additional profit created by said hedge activity.

Just as spreads may widen, or prices may go lower when large selling pressure comes in to the market, remember that the counterparties that were directly involved in the deal may be left with significant spread risk in their book that will need to be managed.  Unexpected changes in price relationships can be exacerbated in the market as a result.

Bottom-line, there is more to this deal than meets the eye.  A simple expectation of lower prices due to hedging activity may not be the only market opportunity!


Click here to receive future analysis

Read More Brynne Kelly Research

Global Oil Spreads Are the Key to Balancing US Oil Inventories

Is a Rotation Out of Oil Into Equities Underway?

Putting Gasoline Inventory Build Into Perspective







Oil Market Favors Bearish Narrative – A Quick Look at This Weeks Statistics

This is an impatient market unwilling to take it’s eye off a bearish narrative and this week’s build in inventory plays right in to that.

Below are a few charts to get a sense of where we are versus this same week in prior years.

Get this Report Delivered to your Inbox each week.

Yearly Comparisons

EIA, Crude Oil, Oil Price Curve, Inventory Comparison

Oil, gasoline and distillate inventories are at similar levels compared to the same inventory week last year, with production quite a bit higher.  By comparison, the entire price curve has shifted $6.00 lower.  How is this year different?

Prior Year June through end of September Inventory Changes

Looking at inventory changes from June – September over the last 3 years, it seems as though the market is bracing for something closer to a 2015 scenario.  This time last year the market was looking forward to some sort of action by OPEC.  Today, the market is eyeing the END of OPEC’s production cuts.

The bearish sentiment this has created is unrelenting and easily fueled by weekly data that once again failed to provide a clear sign that the worst is behind us.  In the last 2 months, total inventories have drawn-down almost 30 million barrels including this week’s build.

We are only a short time into a new era that includes the ability to export oil.  It’s important to remember that the timing of exports from week to week will be volatile. History is still being fashioned.  At this point the market seems unwilling to trust the noise created by the ups and downs of reported exports from one week to the next.

Is There a Case to be Made for Relative Value? Inputs versus Outputs.

by Brynne Kelly

[email protected]

June 2, 2017

Is there a case to be made for relative value?

Selling energy at the beginning of April was a great trade.  However, where do we go from here?  Since prices don’t exist in a vacuum, I thought this would be a good time to take a look at how energy inputs have performed relative to energy outputs.

Think of Crude Oil, Natural Gas and Coal as the three major inputs used to produce energy in a consumable form. Their prices are heavily dependent on the demand and price of the products they can be converted into.  The spread (or margin) between the cost of inputs and the sales price of the outputs not only drives investment in these conversion capabilities, but also indicate supply versus demand.

Get this Report Delivered to your Inbox each week.

With all of the focus this past week on OPEC cuts, inventories and demand worries, let’s take a look at the price movements of the three major inputs.


The entire WTI crude oil curve has fallen more than 6.5%  since the beginning of April, 2017.


The front of the natural gas curve has fallen more than 8% since April.

Meanwhile, CSX coal futures have increased over 6% since the beginning of May, 2017.



To get a feel for the relative value of energy inputs, it makes sense to see how consumable energy products have performed.

First off, we have electricity.  Coal and Natural gas are two of the primary inputs used to produce electricity in the US.  While natural gas prices have fallen significantly in the past two months, the “heat rate” has moved higher.  This means that electricity futures prices are not falling as much as natural gas prices.

Another relative value product to look at is ethane.  While it is not normally thought of as an ‘output’ of natural gas liquids, it does take further processing to get it to market which differentiates it from raw production value.

For the past several years, ethane has routinely been ‘rejected’ into the natural gas stream since the outright price of ethane hasn’t been enough to cover processing costs.  From the chart below, you can see that ethane futures have not experienced nearly the drop that natural gas has this year (at least in the front of the curve).


Another obvious consumable output is gasoline.  Gasoline futures have also fallen less than their underlying ‘input’ falling less than 6% compared to oil’s greater than 6.5% decline.

This relative value play can be seen in the rebound of the gasoline crack from its lows in May.  Meaning that while crude oil prices continued to decline, relative gasoline prices increased.

The same is true for ultra-low sulfur diesel (HO).  As oil prices continued to sell-off, distillate cracks actually started to rebound off their May 1 lows.


What does this mean?  Is this a signal?


Oil Spreads

Another place to spot relative value is in oil spreads.

Relative value of oil in one geographic area to another is similar to input/output economics in that differences in prices can impact the movement of production from one market to another.

Key relationships impacting these movements are US prices versus Europe and Asia.  In the past month, US prices (LLS oil in the gulf coast) through the end of the year have gained value with respect to prices in Europe (Brent), yet European prices have lost value relative to Asia (Dubai).  This is not good for US exports.


However, the spread between US and Asian prices has continued to widen.  Meaning that while the economic incentive to move US oil to Europe hasn’t increased, the incentive to move oil to Asia has increased (as seen in the WTI/Oman spread below).


This points to where the market pins their demand hopes.  As long as global oil demand increases, production to meet that demand can come from anywhere as long as the economics support moving oil from one market to another.

Weekly Inventory

This brings us to the weekly inventory changes.

For the week ending May 26, 2017 total stocks declined by (9.9) per the EIA figures below:

The most important part of this week’s inventory draw is seen in the cumulative change in this “summer” totals versus prior years:

The rate of crude oil stock draw-downs this year relative to the last 3 years is certainly impressive, especially considering that oil inventories posted a net build during this same period in 2016 and 2014.

One contributing factor to this draw-down is the growth in US oil exports:


The chart above definitely highlights what could be considered a ‘trend’ of increasing exports and decreasing net imports.  But, it also highlights the fleeting nature of this dynamic from week-to-week.  However, it’s a positive sign to see these export volumes continue to make new highs.

Speaking of inventory levels, we are approaching that time of year when heating oil (distillate) inventories should start to build in preparation for winter heating season (just as natural gas inventory summer builds in preparation for winter).


With all the focus on gasoline inventory and prices, might it be time to take a look at heating oil prices? Spot prices are half what they were at these same inventory levels 5 years ago.

From a relative-value perspective, there is a lot to think about.

Get this Report Delivered to your Inbox each week.

Read More Brynne Kelly Research

Global Oil Spreads Are the Key to Balancing US Oil Inventories

Is a Rotation Out of Oil Into Equities Underway?

Putting Gasoline Inventory Build Into Perspective

OPEC Cuts And The Evidence To Spot For The Next Market Move

by Brynne Kelly

[email protected]

May 26, 2017

This week the market had to contend not only with weekly inventories, but also with an OPEC meeting and decision.  There was much anticipation and positioning before and immediately after, specifically in the term structure of the curve.  The chart below of the WTI futures curve from May 10 to May 25 highlights this well:

Leading up to the OPEC decision, the curve was decidedly moving into a backward structure starting from around the Dec-17 contract.  The backwardation was at its widest on the 24th (and early the 25th) before news of the meeting hit the wires, but was cut in half after the news was out.

What does this mean and what should you look for in the coming weeks?

Historically there has been a general understanding of what backwardation and contango mean in commodity markets:

Backwardation pulls molecules out of storage.

Contango invites molecules into storage.

Further, Contango invites producers to produce and invest in production while Backwardation leads to future production uncertainty if future prices are unusually low.

Depending on the level of Contango, a producer can cover their cost of carry (storage costs, interest on debt, etc) and then some.  It give producers options regarding how long to ‘carry’ inventory and when and where to deliver their product.  There is comfort in knowing that what you invest in today is worth more in the future (again, depending on the level of the contango).  In addition, it makes it easier to cover the cost of transporting their product on longer journeys.  It takes time to load cargo on a ship and time to get the ship to its final destination.  The level of contango can compensate for that.  A wider contango suggests that the conventional, less-expensive means of storing and transporting have been exhausted and more money is needed to cover the costs of less conventional, more expensive options.

Markets in Backwardation highlight the ‘convenience yield’ of owning a producing asset.  You own a producing asset and can deliver your product immediately into the higher-priced spot market.  Backwardated markets generally reflect an immediate shortage of product not expected to exist in the future.  For example, a major disruption in supply like a pipeline leak will cause the spot market to price at a premium to future markets that reflect the value of the pipeline being fixed.  When a commodity is in demand or fundamentally bullish, the ability to deliver immediately garners a premium (like flowers delivered on Valentine’s Day).  Backwardation can be temporary or sustained (i.e. short-term, weather-induced or a longer-term supply disruption).  Either way, the front of the curve (spot or front-month futures) is the first to react because ‘price’ is the easiest way to balance the change in market conditions.

I mention all of this because the term structures are where people are looking to find and express market sentiment.  The thought was that OPEC would extend their existing cuts to production until sometime in 2018, at which point the cuts would end creating a ‘tighter’ supply demand balance in the short to mid-term than it would experience beyond that.  The expectation being that this might lead to some sort of backwardation in the price curve.  Additionally, one could surmise that backwardation would also help pull molecules out of storage (since it’s more profitable to sell today than in the future) which would organically help to draw down excess inventories.

The problem with this is that it’s more ‘theoretical’ than ‘actual’ at the moment.  Meaning, there isn’t ‘actually’ a shortage of supply in the spot market.  At least not the type of shortage described earlier in the pipeline leak example.  The market continues to keep an eye on gasoline demand with increasing anxiety and the initial impact of an extension of cuts didn’t leave the market with anything tangible to hang its hat on. Rather, it was left with the reality that it will take time for any ‘supply-side’ tightness to actually materialize.  It reminds me of 2014 when US production was materially rising, yet futures prices were reluctant to come off.  It took evidence of increasingly hard-to-find and more expensive storage to finally seal the deal and futures to sell off.  It’s the same this time, there are theories about how these supply cuts might play out and then there will be ‘evidence’.  Look for the evidence.

One place to look for that, obviously,  is in the product markets.  I found it interesting that while crude oil futures across the board shifted materially lower on Thursday, crack spreads were marginally unchanged and have actually moved higher in the last 2 weeks:

The market has built a story around the bearish tone of slowing gasoline demand growth.  Yet the spread margins are not where the selling was.  Does this mean that the issue is not the fixed refining capacity in the US, rather the ability to export incremental barrels?  Meaning we expect there to be enough, however slowing, demand for the amount of finished products the US can both produce for itself and for exports, but that additional barrels of crude can’t economically get to other markets that have more refining capacity.  As discussed in previous posts, global oil spreads have moved slightly in favor of WTI (WTI futures have moved below both Brent and Oman futures) in part, due to OPEC’s earlier production cuts.

Taking a look at oil production versus net US imports since 2010:

and more specifically, in the last 17 months:

Its clear that US net imports did decrease as production was on the rise.  But since the beginning of 2016 we haven’t really gained any ground.  The expectation is that as we would either consume more of our local production and/or export that which we don’t need.  Either way, we need to see a reduction in net imports as proof that current production levels can be sustained without significantly impacting prices.  That is another piece of evidence to look for.

Finally, some comments regarding this week’s inventory numbers.  This week we saw a net draw of (6.1):

I like to put the weekly inventory changes in perspective as a whole by season, like in the Natural Gas markets where inventory is injected into storage over the summer for use (withdrawal) in the winter. While outright inventory levels are important, it’s also key to compare how you are progressing compared to previous years.

In the oil complex, the summer ‘season’ is typically characterized by its demand-related draw-downs of gasoline inventories.  Here’s how this ‘season’ compares to the last 3 years:

If it weren’t for outright inventory levels, especially in gasoline, the seasonal totals so far look fairly impressive.  We clearly have to look for evidence of gasoline demand and/or a reduction in net crude oil imports to have an impact on outright inventory levels.

Using the last 3 year average inventory changes from here through the end of the season (September), this is how inventory levels would stack-up:

Clearly, market bulls would hope for something better than the last 3 year average draw-downs.  This is where the market is hoping OPEC cuts will do the heavy lifting to erode inventories at a faster pace than previous years.  This will only happen if market economics favor exports to move supplies out of the US or if gasoline demand materializes over the summer.

When the market is at a cross-roads, there are a lot of false starts while it waits for something to tip the scale. It’s those that spot the evidence first that catch the real move.