After the price shock: What oil needs to rebound

Inventory-price correlation indicates U.S. crude oil stocks need to shrink to ~450 million barrels for $60 Brent

Source: IEA, BloombergNEF estimates, EIA. Note: model data set from 1Q 2012 to 2Q 2020. *3Q 2020 is quarter-to-date data as of September 17, 2020.

In general, commodity prices have an inverse relationship with inventory levels. Crude oil is no exception.

BloombergNEF created an oil price model using quarterly average U.S. commercial crude inventory data as the independent variable and quarterly average Dated Brent, the European crude benchmark for physical trades, as the dependent variable.

Ideally, we would use OECD crude inventory to estimate Brent prices. However, due to the time-lagged nature of OECD crude stocks data and the high degree of correlation between U.S. and OECD crude stocks, we are using U.S. commercial crude inventory as a proxy for OECD crude inventory.

Brent is a better benchmark for world oil prices and balances because WTI is based on Cushing Hub pricing, and the Cushing Hub has at times been infrastructure-constrained. Using Dated Brent in our model also generated slightly better results than WTI.

Between 2012 and 2Q 2020, there is a reasonable relationship between U.S. crude inventories and quarterly Dated Brent prices.

U.S. crude inventory now stands at 496 million barrels, as of September 11, while the 5-year average is 455 million barrels.
Our inventory-based model indicates that an inventory of around 450 million barrels would signal Dated Brent to trade closer to $60/bbl.

Actual Dated Brent prices for 3Q 2020 (quarter-to-date) is averaging above what our inventory-based model indicated (see green dot on chart). However, the difference is well within the historical margin of error. Moreover, if we average out 1Q 2020 to 3Q 2020, the modeled results of $43.57 is fairly close to the actual of $41.83.

China slowdown, global pandemic caused this historical relationship to break in the past

For reality check, we took a closer look at the historical quarters in which U.S. crude stocks were near 450 million barrels and checked the corresponding average Dated Brent prices.

Since 2012, Dated Brent often averaged slightly above $60 per barrel during these quarters.

There were two exceptions:

In 4Q 2015, Dated Brent averaged $42 and U.S. crude stocks stood at 454 million barrels. We attribute this to demand worries. The Chinese economy had slowed to a sub-7% rate growth rate year-on-year in 3Q 2015, something the world had not seen since the 2008 global recession. Investors were fearful that crude demand would slow along with China’s economy.

In 1Q 2020, Dated Brent prices also averaged well below what our inventory-based model suggests. This time, a Saudi-Russian crude price war and pandemic-led demand destruction devastated crude oil prices.

Note: It would be unrealistic to rely entirely on our inventory-based model to estimate Dated Brent prices. After all, crude oil is affected by other factors such as geopolitical events, expectations of world economy performance, investor sentiments, weather events, as well as other fundamental factors. This modeling effort is intended to help lend insight into how a high-frequency data point, namely storage, can provide a bellwether to broader market trends.