Thursday, January 26, 2017

Crude correlations about crude oil can shed light on changes

Statistical correlations are interesting things.

One of my favorite websites, and now a book, is Spurious Correlations. The site’s author looks at seemingly unrelated data sets that correlate in interesting ways. For example, there is a high degree of statistical correlation between the per capita consumption of chicken and total US crude imports.

Or, my personal favorite, you can look at the correlation between the annual number of swimming pool drownings in the US and the number of films actor Nicholas Cage has appeared in.

Number of people who drowned by falling into a pool correlates with films Nicholas Cage appeared in

Unsurprisingly, there’s also a high degree of correlation between various regional and global crude oil crude benchmarks. And while correlation doesn’t equal causation, those relationships have changed in interesting ways over the past year.

Historically, there was a strong correlation between the cash differentials for Light Louisiana Sweet crude and the Brent-WTI spread. Since restrictions around US crude exports have been lifted, those statistical relationships have just about vanished.

Looking back to 2012, on average there was an 85% correlation between the differential between LLS and WTI and the Brent-WTI spread, meaning that for the most part when the Brent-WTI spread widened, the LLS-WTI spread also widened.

This relationship made a lot of sense. At the time, WTI was landlocked in Cushing, Oklahoma, and LLS reflected the market for light sweet crude among the Gulf Coast refiners, which account for more than half of total US refining capacity.

Gulf Coast refiners looking at a light sweet barrel had a few options: either buy LLS, or a related domestic grade, or import a similar barrel that would almost certainly be priced in relation to Brent crude, which was generally more reflective of supply and demand in the international crude market.

At time, Brent was traded $15-$20/b above WTI, which was essentially limited to the US Midwest refining market. At the time, Cushing was oversupplied with domestic and Canadian crude, which had no way to reach refiners in the Gulf Coast.

So it made a lot of sense for LLS to fetch a market price much closer to Brent than WTI. Refiners were looking to pay Brent-level prices for import crudes, why would domestic sellers offer LLS at $15 under that market? That’s how markets work.

The correlation between LLS and Brent diverged somewhat in 2013 and 2015, as new pipeline projects came online, providing access for Gulf Coast refiners to more light sweet crude from West Texas and Cushing.  The Seaway Pipeline reversal project, which converted crude line to carry oil from Cushing to Freeport, Texas, helped drive the WTI-Brent spread back under $10/b. In 2015, the correlation between the two averaged at 84%.

Then, in December 2015, the US announced change to its crude export policy. With US crude now freely exportable, the correlation 2016 dropped to average just 1%.

Correlations among US crude assessments
In 2016, the Brent-WTI spread averaged $1.35/b, compared to $4.67 in 2015 and $7.05/b in 2014. This is one logical outcome of the change to US export policy.

If the WTI-Brent spread becomes too wide, then US producers will find international buyers for cheaper-WTI linked crude (like LLS). At the same time, with more than 1.5 million b/d of pipeline capacity from Cushing and West Texas to the Gulf Coast, there are plenty of light, sweet alternatives to LLS should the differential widen out again. This keeps essentially provides something of a ceiling on both LLS and the Brent-WTI spread.

Correlation doesn’t equal causation, but it can shed some light on the relationships that make up these markets.

No comments:

Post a Comment