As you’re likely aware, crude oil inventories in OECD nations can be a strong indicator of the directionality of total global inventories. Examining OECD inventories can help anticipate market balances and therefore potential price movements.
By using Ursa Europe Oil Storage data and other trusted sources, we can monitor what has occurred in Q1 of 2018 for OECD inventories to gain this strong understanding – and the competitive advantages it offers.
The International Energy Agency (IEA) breaks down OECD inventories into 3 major groupings: 1) OECD Americas, 2) OECD Europe, and 3) OECD Asia Oceania. For each of these regions, we can use a different data set to accurately represent total crude inventory levels changes:
OECD Americas – U.S. inventory data from the Energy Information Administration (EIA) is a trusted source to give total U.S. crude inventory levels, and the U.S. has almost a 1:1 relationship with OECD Americas from the IEA, meaning simply looking at EIA weekly data on the U.S. provides a proxy for total.
OECD Europe – There is currently no good source of weekly European inventory levels, and IEA data provides total figures 2-3 months delayed, so we will substitute Ursa European Oil Storage levels here.
OECD Asia Oceania – Japanese inventories make up a significant portion of OECD Asia Oceania, and the Petroleum Association of Japan (PAJ) provides trusted weekly levels of crude oil in Japan. Weekly PAJ data can be used as a proxy for total OECD Asia Oceania.
To mathematically demonstrate #2 above (Ursa data substituted for Europe totals), we define 6 key countries that have significant crude inventories that Ursa covers a significant portion of (A.K.A. Ursa6). We then demonstrate that the Ursa6 can be used as a proxy for total OECD Europe inventories. We use a long history of IEA data on these Ursa6 compared to total IEA OECD Europe to show the statistical relationship in the following regression model:
The R² of 0.710 value tells us that Ursa6 country data from the IEA captures the majority of the variation that is seen in total OECD Europe data. Thus, Ursa’s more recent Q1 European data can be used to substituted total OECD Europe data, to at least demonstrate change during the quarter.
Now combining Ursa6 data now provided by Ursa with EIA and PAJ data for the U.S. and Japan respectively, we can examine global OECD inventories.
The following chart below shows our Total OECD proxy for Q1, 2018:
Clearly, we are seeing builds over Q1 in inventories. Over this time period, this data (Ursa + EIA + PAJ) shows that inventory builds are up about 4%, or ~24 Mbbls over Q1. The average Q1 builds over the past five years using IEA Total OECD data is about 6%. We see that in 2018, there is less than typical builds based on seasonality. This suggests the potential impact of supply cuts or an increased demand.
We see primarily that these builds are due to U.S. (12.7 Mbbls) and European stock builds (11.5 Mbbls). Japanese data stayed relatively flat in the period (0.1 Mbbls).
As is seen in the chart above, January and February showed the largest build, while March appeared to stay relatively constant. We can dive a bit deeper into Jan/Feb:
The inventory builds from early January to end of February are up 4% according to the Ursa proxy, and the average IEA Total OECD data for the past five years show builds of 5%. There is a small discrepancy between Ursa data trends for this year and IEA January-February trends for the past five years, indicating that builds are certainly occurring, but a bit slower than usual.
Much like January and February, the Ursa proxy for March trends with, but does not precisely match, what is typical of this month over the past five years according to IEA Total OECD data. The Ursa proxy reports negligible, near zero, change in inventories, while the average IEA build over the past 5 years is 1%. Because the IEA still does not have March data and the combination of Ursa, IEA and PAJ data match the IEA 2018 Jan-Feb numbers, we anticipate that we will see normal seasonal trends in inventories for March.
So what does all of this tell us? Well, to start, our Ursa proxy matches what the IEA saw in OECD builds for January and February of this year, and the Ursa proxy indicates that Q1 builds are lower than usual. Because the IEA does not discuss March data yet, we suspect with our estimates that the IEA will report that there will be minimal change in Total OECD inventories for the month of March. As previously mentioned, the slightly lower builds could be a factor of either limited supply or increased demand.
Given that Ursa provides weekly data, it is possible to see inventory trends well before the IEA data is published without any lag, and make oil market inferences based on this data. By combining other trusted sources of inventory data with Ursa, we demonstrate a full OECD picture that potentially predicts future IEA analysis.