Stock take

• Oil inventories are one of the preferred targets of oil market policymakers as stockpiles neatly aggregate all the varying supply and demand variables into a single figure. We estimate that OECD stockpiles—those in developed markets—ended 2020 at a little under 3.1bn bbl, around 190m bbl higher than a year earlier. However, there was wide movement over the course of last year as a slump in demand related to the Covid-19 pandemic and price war activity saw an enormous build of 325m bbl in Q2 alone.

Oil inventories rose dramatically in 2020

Source: IEA, Emirates NBD Research

• While still a large figure in and of itself, the overall level of oil and product stocks tells us little without the context of measuring their movement over time or against how well inventories can serve demand. Mohammed Barkindo, OPEC’s secretary general, has repeatedly highlighted bringing OECD oil stockpiles down to their five-year average as an objective of the OPEC+ production cuts, noting that inventories were still some 160m bbl above that time-based target. However, in our view movements up and down in stock levels are only partially helpful in determining the relative condition of oil markets.
• A prolonged period of inventories increasing could indicate that supply is outpacing demand and thus markets are loose or vice versa, persistent declines in inventories could indicate that demand is rising and markets are tight. But we believe that measuring stockpiles against demand is a more powerful indicator in determining whether oil markets are tight or loose and has a meaningful relationship on the short-term direction of prices.
• Targeting a five-year average of inventories fails to account for underlying changes in supply and demand. Oil markets over the last five years have generally seen inventories rise as the market struggled to absorb rising production from the US, Canada, Brazil and elsewhere and masks extraneous dynamics affecting the market, such as the Covid-19 pandemic and the OPEC-Russia price war in early 2020.

Supply has been outpacing demand over the last five years

Source: IEA, Emirates NBD Research. Note: global oil supply and demand levels.

• Our preferred measure for oil inventories is how well they satisfy demand conditions over a given period. The IEA obligates its members to hold 90 days worth of net oil imports as reserves which can be met from either publicly held strategic stockpiles or via stocks used for commercial purposes. Over the last 10 years or so that has meant that commercial stockpiles have averaged around 60 days worth of demand in OECD countries.

OECD stocks have historically averaged around 60 days worth of demand

Source: IEA, Emirates NBD Research. Note: commercial stocks only.

• Prolonged periods of stability at around that 60 day level appear to be associated with stability in oil prices while divergences away from that level appear to have a meaningful relationship with oil price volatility. During the 2014-16 boom/bust cycle of shale oil and OPEC’s market share response, OECD stocks rose to a peak of around 67 days worth of demand before eventually normalizing. In 2020, OECD stocks peaked at more than 85 days worth of forward demand.
• Rather than targeting a five-year average of overall OECD stocks we find that there is a meaningful relationship between oil prices and a shorter-term target of a rolling two-year average of days of demand. Movements above that two-year average imply lower oil prices and vice versa. Indeed, applying the model to y/y movements in Brent futures (over a quarterly average basis) seems to fit the performance of prices better than targeting returning OECD stocks to their rolling five-year average level.

Days of demand model vs five-year average level model

Source: Bloomberg, Emirates NBD Research.

• The outlier move in Q2 2020 when the days of demand model was some 300% below the actual price performance is worth highlighting given that it over-exaggerated the impact of the enormous build in stocks during that period. However, as a signal to expect heightened downside volatility we believe the model has merit: Brent futures did collapse during the start of Q2, bottoming out at around USD 16/b in mid-April while WTI futures went negative. 
• We expect to see global oil inventories draw on average this year, by around 1.2m b/d. That should help to bring the ‘excess’ days of demand over their two-year average down and indeed negative in a considerable way by the second half of the year. Our model suggests then that Brent prices can rally sustainably provided those stock draws remain in place. Indeed, the model’s price output at an average of USD 50.48/b is not far off our qualitative assumption that Brent prices will record an average of USD 50/b.