Contango vs backwardation: what the curve shape means for storage economics

Every desk has run this math at some point: the front month is trading under the second month, storage looks free, so why isn't everyone filling tanks? The answer sits in the gap between what you pay to carry a barrel and what the curve pays you to hold it. That gap, cost of carry measured against the forward curve, is the whole trade.

Contango pays you to wait, backwardation doesn't

Contango is when forward prices sit above spot, each month out trading a little richer than the one before it. That shape rewards storage: buy the barrel now, sell it forward, and the spread between the two covers tank rent, financing, and insurance with something left over. The 2020 WTI collapse, when May traded negative and the curve blew out into steep contango, is the case everyone remembers. Floating storage filled fast because the curve was paying absurd money just to sit on barrels.

Backwardation flips that. Spot trades above forward, so the market is telling you the barrel is worth more today than it will be worth later. Nobody gets paid to store in that structure. Storage still happens operationally, you need working stock, pipeline fill, strategic reserves, but the incentive to add incremental barrels into tankage disappears. When a market is backwardated, draws on inventory tend to run ahead of what fundamentals alone would suggest, because the curve itself is pushing barrels out of storage and into the physical market.

Running the cost of carry math

The trade only works if the contango spread exceeds your all-in carry cost. That cost has three components a desk prices: the tank or vessel day rate, the interest on capital tied up in the barrel, and insurance plus any quality loss over the hold period. Add those up and compare against the spread between the purchase month and the sale month. If the spread clears that number, the contango-storage trade is live. If it doesn't, the curve shape is interesting but not actionable.

Floating storage economics add a wrinkle onshore tanks don't have. A VLCC chartered as a floating tank runs a far higher day rate than shore tankage, so the contango has to be considerably wider before putting vessels on charter for storage pencils out. That's why floating storage tends to show up late in a contango cycle, after onshore tank farms are already tight on space. Land-based capacity is the cheap seat; ships are the overflow valve once that seat is gone.

Why the curve and the physical count don't always agree

The curve tells you the incentive, not the inventory level. A deep contango says barrels should be going into storage. It doesn't say how fast tanks are filling, or whether a given hub is already near tank-tops while another sits half empty. Agency reports land once a week and reflect data that's already a few days old by the time it's published, so a desk pricing the roll is often working off last week's picture while guessing at this week's draw or build.

That gap between what the curve implies and what's actually in the tanks is where an independent, daily stocks read earns its keep. Tracking floating roof position by satellite pass, rather than waiting on self-reported terminal data, gives a desk a way to check whether the physical fill is tracking the contango-driven incentive or running ahead or behind it. Oil Inventory Index builds that signal from the shadow a floating roof casts, delivered as a daily series per facility or region, so the curve story and the tank story can be checked against each other before the weekly number lands.

If you're trying to trade the roll instead of just watching it, a daily independent stocks read is the piece most desks are still missing.