Understanding differentials and basis risk

A basis spread, or differential, is simply the difference between the actual price paid or received for oil or gas of a specific quality grade or at a specific delivery location, compared to the benchmark oil or gas price.

Differences between prices give us clues about relative abundance or scarcity in different markets, and allow markets to find an equilibrium by redistributing supplies or incentivizing alternatives in response to those price spreads. There are so many different grades, qualities, and delivery locations of energy products that it would be impossible for all of them to trade as equally active, liquid markets. A healthy market has a degree of liquidity and fungibility, so identifying a relatively small number of well-defined and transparent benchmarks helps energy markets function well. Less liquid energy commodities can then price as differentials to those benchmarks.

Broadly speaking, for crude oil and refined oil products, differentials are dictated by quality and delivery location. For natural gas, delivery location is really what matters. Although natural gas can have different qualities at the wellhead (mainly due to differences in the percentage of natural gas liquids in the stream), after processing, “pipe gas” is pretty much all the same. Another way of thinking about delivery location is in terms of transportation cost — how much does it cost to get crude, refined product, or gas from the place it was produced to the delivery hub where it will find its buyer? That cost generally would represent a minimum price spread.

Basis risk is the idea that a producer or consumer referencing benchmark pricing as a proxy for their specific price exposure do face residual uncertainty around exactly what they will pay or receive for their energy. For example, the price difference between West Texas Intermediate (WTI) crude delivered to Midland, Texas and WTI delivered to Cushing, Oklahoma (which is the delivery location for the Chicago Mercantile Exchange traded US crude benchmark) would be the basis risk of an oil producer in the West Texas Permian Basin selling outside of that region. Similarly, the difference between the Appalachian natural gas price and benchmark, exchange-traded Henry Hub gas would be the basis risk of a producer of Appalachian gas. The idea applies to refined products, too — the difference between a local jet fuel acquisition price and a benchmark diesel price, for example, would represent an airline’s basis risk if the company hedged, or agreed a long-term purchase contract, based on the benchmark.

For energy products with a high and consistent degree of correlation to a benchmark, producers and consumers don’t necessarily have to worry too much about basis risk. Where basis risk becomes more important — to be aware of, and to try to manage — is when correlation to the benchmark is low or erratic.

What determines crude oil quality differentials?

Crude oil is not homogenous. There are hundreds of different grades that have different characteristics. The two most important characteristics are:

• API gravity, indicating how heavy or light a grade is, where a higher value indicates a lighter crude; and

• Sulfur content, indicating how sweet or sour a grade is, where a higher value indicates a more sour, higher sulfur content crude.

Generally speaking, lighter crudes also tend to be sweeter (i.e. have a lower sulfur content); heavier crudes tend to be sourer. But this is a very rough correlation. The chart below shows a selection of the world’s major global crude grades arranged by quality; the size of each circle indicates the relative volume currently produced. Grades near the top, left of the chart are light, sweet grades. Crudes near the bottom, right of the chart are heavy, sour crudes. Roll over each region to see just the crudes in that region. North Sea, West African and Asian crudes are all relatively light and sweet. Russian crudes are typically middle grades. Middle Eastern and Latin American crudes run the gamut but are typically more skewed to the heavy, sour end of the range. Notably, there is a “barbell” distribution of crude qualities in North America — the highest volume US grades are very light and sweet, while the highest volume Canadian and Mexican grades are heavy, sour.

These characteristics make different crude grades more or less desirable to different refiners because, in very simplistic terms, you get out of a refinery what you put in. Different crudes yield different percentages of each type of refined product or require more or less processing to do so. A simple refinery needs a “high quality” crude — a light, sweet crude — to yield a high-value basket of products, like gasoline and diesel. A complex, sophisticated refinery has invested in upgrading units that allow it to take a lower quality barrel of crude — a heavy, sour barrel — and still yield a high-quality slate of refined products through additional steps of processing. Historically, in general, heavy, sour crude priced at a discount to light, sweet crude for this reason (more recently this relationship has become more complicated and this is discussed in more detail below).

There are also other less cited crude oil characteristics, like viscosity, boiling point, or the content of asphaltines, resins, or nitrogen. A crude assay is a full assessment of the chemical composition of a crude. This can change over the life of an oil field, so will generally be repeated periodically. Broadly speaking all of these metrics together give an indication of what a refiner can expect to get out of a barrel of crude, given the refiner’s configuration. A refiner must consider a crude grade’s characteristics in totality when making a decision about what type of crude to process and how much to pay for it, and this becomes more complicated when a refiner is blending different types of crude together.

The reason this is important is that when there is a disruption in crude supply, it is important to understand not only how big the disruption was in volume terms, but what quality of crude was disrupted. This will be reflected in crude differentials, and possibly also refined product spreads to crude and regional refined product spreads. Relative pricing should give an indication of exactly which refiners will be most affected by the disruption, and how easy or difficult it might be to replace the lost barrels with a substitute. For example, if two light sweet crude oil production streams are interrupted at the same time in a given region, that might have a different market impact than if one light stream and one heavy stream were interrupted at the same time, even if the overall volume disruption was similar.

What determines refined product quality differentials?

Refined products are relatively more uniform than crude in terms of their quality, but still have some variation. The most common type of quality difference in refined products has to do with sulfur content — diesel and residual fuel oil can both vary in terms of sulfur content, though increasingly international laws have harmonized over time to allow less and less sulfur in refined products in most locations. Gasoline also has different grades that reflect octane level.

Country-level and international regulations of refined product specifications change over time, and those changes in regulations in and of themselves can have a major impact on expected and/or actual refined product pricing and differentials before and during the period of implementation. One recent example of a change in refined product regulations is the IMO (International Maritime Organization) regulation of sulphur oxide emissions from ships, which started in 2005 and has become progressively more stringent. Between 2005-2020 the rules gradually phased down the allowable sulphur content of the fuel used by ocean-going ships, with the biggest reduction (from 3.5% sulphur by weight to 0.5%) in effect as of January, 2020. “IMO 2020” was widely anticipated to impact refined product differentials (to the point of affecting even underlying crude pricing) as shippers phased out the use of higher sulphur residual fuel oil, or bunker fuel, in favor of lower sulphur gas oil or diesel.

How is locational, or geographic basis risk determined for oil and gas?

Locational basis risk mainly comes down to the fact that the most likely buyer of an energy product is not necessarily in the same place as where that energy product is produced. There is some explicit transport cost associated with getting the right energy product to the right consumer at the right time, and there may also be an implicit cost of incentivizing that movement from one place to another in the first place. Three main things can change locational price spreads:

• Changes in relative demand in different locations;

• Changes in relative supply or production in different locations; and

• Changes in transportation availability or costs.

Starting with the example of crude oil: A producer of crude oil ultimately sells to a refiner. It is possible that the refiner is “down the road”, so to speak, from the wellhead, in which case the producer and refiner will reference the same price location, and neither has to worry about locational basis risk. But crude that is produced in the Permian region of West Texas, for example, will find a lot more refiners in the Houston area, or even the US Midcontinent that would have an interest in buying that crude. For that reason that crude will have more value in Houston, Oklahoma or Chicago, and less value, most of the time, in West Texas. That price signal “pulls” crude out of West Texas, and the magnitude of that price difference depends mainly on how difficult (and expensive) it is to transport the crude interregionally.

A very efficient market, with ample distributional infrastructure, will have a very consistent basis spread to the benchmark in the absence of unexpected disruptions to supply or demand, and that spread will reflect the relevant pipeline tariff or shipping cost. So, in other words, “normal” supply and demand patterns dictate a “normal” direction of flow from one region to another, and if pipeline capacity from a production basin to multiple refining centers is sufficient and the price of that transportation is always about the same, the differential between the regional crude grade and the benchmark crude price should be fairly consistent and roughly equal to the shipping cost. On the other hand, if it is very difficult or expensive to move crude out of the production basin to one or more major markets — because there is not enough distributional infrastructure, not enough optionality in where that infrastructure goes, or because that infrastructure is frequently interrupted — then there is a risk of the differential between the regional crude grade and the benchmark being much wider and/or more erratic.

The idea for natural gas is the same. To the degree that local consumption is not equal to local production, gas has to move from one place to another (overwhelmingly by pipeline in North America). In theory, pipeline capacity is designed in anticipation of those flows. For routes on which there is plenty of pipeline capacity, the basis spread between the production hub and Henry Hub should be pretty consistent and equal to the pipeline tariff. To the degree that pipeline capacity on a route is sometimes or frequently inadequate, the basis spread will be wide and/or erratic. On some routes, there may be an inherent seasonality in the magnitude of regional price spreads due to seasonal variation in inter-regional demand. And in some cases, due to rapid changes in regional supply or demand for natural gas, pipeline capacity may be structurally inadequate for some period of time (examples of this are discussed below).

The idea for refined oil products is also basically the same — the difference between the price of diesel in the US Gulf versus New York Harbor, or the difference between the price in Northwest Europe versus the Mediterranean, signals relative scarcity or abundance. All else equal, the price spread should reflect the cost of transportation between markets. But in more exaggerated situations — extreme or unexpected abundance or scarcity in one region relative to another — spreads will widen to signal the market to move products around until the equilibrium is regained and the spread is once again reduced to the typical cost of transport.

Examples of structural changes in locational & quality price differentials

There are many recent examples of massive widening in locational differentials for North American crude and natural gas as a result of unexpected changes in infrastructure requirements. The recent boom in US oil and gas production has occurred primarily in regions that did not previously have a lot of production. For oil, that boom happened first in the Bakken (North Dakota) then the Permian (West Texas). The boom in natural gas production started in and has been dominated by the Appalachian region of Pennsylvania and surrounding states.

But prior to the shale/tight oil boom, the US Gulf was the most important US oil and gas production area. The US was also much more reliant on waterborne imports of oil and the expectation was that dependence on waterborne imports of natural gas in the form of liquefied natural gas (LNG) would also grow. Reflecting this old paradigm, the US oil pipeline network was primarily designed to move barrels from the US Gulf north to Mid-continent refiners (US Gulf refiners used local domestic supply or waterborne imports, and East and West Coast refineries relied on waterborne cargos as well). The natural gas pipeline network moved supplies primarily from the US Gulf up into the Northeast, moved Western Canadian gas supplies south and east, and later moved gas production from the US Rockies region east as well.

As a result, when new oil and gas production started up — rapidly and in very large volumes — in areas where there had not previously been much if any production, and where local demand was limited, takeaway capacity for those volumes simply didn’t exist. This is why, at various points in time, we have seen massive discounts in the prices of “stranded” Bakken and Permian crude, and Appalachian gas; without a way to reach a buyer, the market value of these supplies was limited when they were first developed.

Price spreads for oil and gas supplies from new production areas have also been very erratic as production growth has accelerated and slowed, and as new infrastructure additions have caught up to supply, then fallen behind again. The market always does its job — massive price spreads always eventually do incentivize new takeaway capacity – but it does not happen overnight and happens in fits and starts. And as new sources of US oil and gas production did make their way to market hubs, there was a second order impact on other, traditional supply sources that were displaced, such as Western Canadian oil and gas. As these supply sources were squeezed out of their traditional markets by new supply sources, that reduced demand was reflected in steep price discounts to benchmark prices.

The rapid growth in US tight oil production also changed some old rules of thumbs about crude quality differentials. Historically, heavy, sour crude almost always traded at a discount to light, sweet crude. In response, about 10-20 years ago many US (and global) refiners invested a lot of money in upgrading units that allowed them to take advantage of a structural price discount by running heavy, sour crude while still maintaining a high yield of quality refined products.  But the tight oil boom increased production of very light, sweet crude specifically, at a time when that was no longer the preference of most US refiners. As a result, the relationship between the prices of different crude qualities has also changed.

Early in the boom, a significant increase in Bakken crude supply into the US Mid-continent (and a dirth of exit points from the Mid-continent other than to local refineries, which were configured to run heavy, sour grades) put tremendous pressure on WTI-Cushing prices relative to other crudes. This was an unusual situation, as WTI-Cushing is the exchange-traded North American benchmark, but suddenly was neither representative nor fungible. Other grades — such as LLS, a light sweet crude in the US Gulf — traded at big price premiums to WTI-Cushing. This was not a function of the fundamentals of those other grades though, it was a function of fundamentals that were very particular to WTI-Cushing.

Some changes in price spreads are temporary or situational 

A temporary, unexpected interruption in a source of supply of a certain quality of crude — from a storm, a technical glitch, or a geopolitical event — can be reflected in quality differentials across several grades. For example, a sudden interruption in large volumes of heavy, sour Venezuelan crude supply would likely cause all crude prices — including even benchmark light, sweet WTI — to go up. But because Venezuelan crude is heavy, we would almost certainly see prices for US Gulf heavy crude grades go up even more than WTI — in other words, all crude prices would rise, but the heavy discount to the benchmark would narrow.

Similarly, a wellhead shut-in, a pipeline disruption, or a processing facility disruption could have a major impact on a certain natural gas basis spread, and may or may not be significant enough to be reflected in the benchmark Henry Hub price as well. On the refined oil products side, a big increase or decrease in airline activity in one region compared to another, for example, could be reflected in the spread between jet fuel prices in those two regions. And a big increase in shipping rates on a certain tanker route could in theory affect the relative pricing of crude at either end of that route. Many supply or demand events that affect oil and gas — if significant enough — affect benchmark prices too. They just affect certain quality grades or locations even more. 

How can or should basis risk be managed?

As with any other risk management program, hedging cannot structurally increase or decrease one’s basis spread or differential, at least not over a long period of time. What it can do is smooth out variation in the spread. There is often (though not always) a trade-off between specificity of risk, and liquidity. Major benchmarks have the best liquidity, but relatively few hedgers have specific price exposure to those commodities — in other words, relatively few producers or consumers actually physically produce or consume the benchmark products such as WTI or Henry Hub. However, there might be a liquidity premium — a higher hedging cost — for hedging specific risk exposure (to, say, Bakken crude, Appalachian gas, or Singaporean jet fuel) with less liquid instruments.

Basis hedging is most important is when basis risk is highly variable. Suppose, for example, a crude oil producer is interested in hedging price exposure and that the crude grade that is produced – the producer’s specific price exposure – trades at a wide, 20% discount to the benchmark crude. That is not exactly a problem (or at least not one that can be fixed by hedging) so long as that discount is consistent. Knowing that the discount is always about 20%, that producer can hedge overall price direction with the benchmark grade, and have a reasonable expectation about future revenues. However, if the discount varies widely then hedging with the benchmark leaves unmitigated price exposure.

Looked at another way, if the price of the produced grade has an 80-90% correlation with the benchmark price, that producer has effectively hedged 80-90% of their price exposure by hedging with the benchmark rather than with the grade they actually produce. But if the price of the produced grade has only a 30-40% correlation with the benchmark, a much bigger portion of the producer’s price exposure is floating or unhedged.

The producer might decide that it is worthwhile to pay the “liquidity premium” associated with hedging with a less liquid but more correlated grade. Alternatively, the producer may decide to use the benchmark to hedge the directional price risk associated with most of its production volume, but to layer over that a smaller volume hedge in a more correlated grade. Except in cases where specific risk exposure happens to be to a benchmark product, the decision is between correlation and liquidity.

What’s the best way to keep an eye on basis spreads or differentials?

When you observe a change in the benchmark oil or gas price and are looking for an explanation, it can be helpful to look at a few basis markets, too, to see if the benchmark is leading the price move or following an even bigger price move in another grade or hub. Similarly, changes in the prices of individual grades or basis locations can often be an early warning sign about a move in the benchmark price. For example, price changes in certain North Sea grades can foreshadow price changes in exchange-traded Brent.

Pick a few key price spreads to watch over time that reflect major trading locations and (for oil) a range of qualities. Track these relationships as spreads over time, and as correlations over time. Bear in mind that, as a general rule of thumb, spreads tend to widen when price level goes up, and narrow when price level goes down. In other words, let’s say that a given crude grade usually trades a discount to WTI. The higher WTI prices go, the bigger that grade’s discount is likely to be. As WTI prices fall, the spread is likely to narrow. For this reason, it is useful to watch spreads over time not only in dollar terms, but in terms of the grade’s price as a percentage of the benchmark price as well.