As the oil market collapse has taken WTI into the uncharted and “impossible” territory of negative prices, there has been considerable attention not only on where prices might be going, but also what is happening to the forward curve, and what it means. Even in more normal markets, the structure of the forward curve “backwardation” and “contango” is a source of confusion – mystery even.
Some of the comments made as the markets have plunged – and the market has gone into what can accurately be described as “super-contango” – have not necessarily helped and at times have fed common misconceptions. The following discourse is intended to give a more accessible explanation.
When attempting to value – an important word here – an asset at a future date (i.e. a price today, for delivery at a future date), commodity markets pose a particular problem. In other markets, such as foreign exchange, given the spot rate and the interest rates in the two currencies, the forward rate defines itself: as you can buy dollars today and keep them on deposit for 12 months as freely as you can hold Sterling on deposit and buy dollars in 12 months’ time, if the forward rate isn’t the spot rate times the ratio of the interest rates, there would be a risk free arbitrage going one way and coming back the other. Markets don’t let that sort of thing happen.
In contrast, commodity markets don’t offer any such mechanism. So, if the price of oil is $50 (for those of you with long memories) today, what should the price for delivery in a week’s or a month’s or a year’s time be? The answer, as it does for all manner of price relationships in commodity markets, lies in supply and demand.
In a market that is generally perceived as balanced – there’s enough to go round, but not an excess, and that is expected to continue – there is no reason why oil next week or next month shouldn’t also be $50.
However, if there is a supply disruption, that will set buyers scrambling. And those buyers will cover a range from those who are thinking they need something soon to those who definitely need it now. The market will obviously spike upwards (more buyers than sellers) but the key is that the buyer who needs product now has to pay a premium over the buyer who only needs it next week to ensure he gets his supply. The inevitable and necessary consequence is a backwardation – where the prompt price trades at a premium to the forward prices. It is a clear signal of market strength, and prices are expected to rise. If the supply shortfall persists, prices will continue to rise, and the backwardation can be expected to steepen.
Contango is generated in an analogous, if not quite identical way. In an oversupplied market, a supplier whose stocks are building because of a dip in demand needs to discount his product to encourage someone who doesn’t yet need to buy to do so, or face the consequences of stopping production. Those who are going to be at tank tops tomorrow obviously need to discount more than those who are a week to away from being full. The inducement to the discretionary buyer is reflected in the spot price trading at a discount to the forward – contango. Again, if the supply surplus persists, prices will continue to fall, and the contango steepen – perhaps? Let’s come back to that shortly.
But in the behavioural mechanism generating the forward structure, there are three key things to recognise. First, that the valuation happening is relative not absolute. The question is how much premium is needed to keep next week buyer’s hands off – or how much discount to persuade a buyer who doesn’t really need to buy to play. Absolute price then turns out to be more of a consequence. While the market will use the last price as a guide, a starting point even, more important than where the last seller (or buyer) was is where the next one is. In a tight market, the next traded price will be higher whether the market is at $10 or $100.
Secondly, the market action is all driven from the front. The forward curve doesn’t really evolve from the spot price; instead, the spot price reacts relative to a more stable forward price, with backwardation forming because the spot price rises or contango because the spot price falls.
Thirdly, and perhaps the most important in understanding forward structure, the mechanism generating this forward structure has no predictive or forecast component. In an article published by one of the major accountancy firms just last week, one of their justifications for having a forecast above the forward curve was that “the forward curve is a very poor predictor of price”. The point they should have made is that the forward curve is not a predictor of price at all. The June futures price is today’s valuation for taking delivery in June relative to any other time; it’s a bit of a stretch to expect it also to be the forecast for where the market will be when we get there.
There is – at least in normal circumstances – one critical difference between backwardation and contango. In a backwardated market, there is no mechanism – no safety-valve – available to limit the amount of backwardation a market can sustain. Essentially there is no way to deliver today product that doesn’t exist until tomorrow. Someone who really needs it will pay just about any premium to get hold of product because the consequence of not doing so is beyond contemplation: such as grounding a fleet and facing the immeasurable consequences.
In contango, however, it is possible to buy something today you don’t need until tomorrow, as long as you’ve got somewhere to store it. Consequently, the distressed seller generally only has to discount enough to make it worth taking into storage, all because the market will nickel and dime any arbitrage opportunity for a profit. When competing for a free lunch, there is always a trader who will settle for a free breakfast. As a result, the contango will tend to be limited to the prevailing cost of storage.
Not all storage is the same. Not only are there different types of tanks for different usages, including using tankers as floating storage, each with their own cost, they also have different owners with different economics around using them. And there is a market for storage, with storage costs rising and falling with demand (not surprisingly, well correlated with whether the market is in backwardation or contango). But generally, in a more orderly market decline, it is possible to see the levels of contango increase as first the cheapest and then the more costly forms of storage are filled.
Where it breaks down, and where the contango can go unlimited – “super-contango” – like backwardation is where storage runs out. And that is essentially what we have just witnessed – and a key feature is that it is locally driven as well. The particular issue with WTI is that its delivery point is landlocked Cushing in Oklahoma. If the oil is being is produced it has to go somewhere, and if there is nowhere for it to go, someone has to pay “Mr Fix-it” to sort things out. The oil market has suddenly, if briefly, taken on the unstorability of an electricity market. Essentially, Mr Fix-it’s fee has turned up in the discount for prompt delivery versus the forward, and produced a negative price.
It is unlikely we will experience negative prices in Brent. That it has happened with WTI only serves to highlight the issue of selecting a landlocked hub susceptible to localised distortions as a supposedly universal benchmark. Brent’s multiple-source, open-water delivered basket of crudes should mean it avoids that fate, but it will still be able to exhibit severe levels of contango.
The Golden rule that prices can’t go negative has seemingly been broken. Economically the rule is still true to the extent that negative prices are unsustainable. But practical expediency, pre-existing commitment, and inability to stop everything instantaneously are all likely to have played their roles in allowing prices to go negative. Ultimately, where OPEC+ have failed, the market will succeed. It has to. If the current conditions persist someone will have to blink, and production will be shut in. The ultimate storage – keeping it in the ground – will be the last resort. And the contango will persist until the market restores supply balance, and can start eating into the accumulated stockpiles, even if the outright market price recovers.
The information the market forward curves hold is very important. Arguably, the market structure gives a better indication of the state of the market than outright price or recent activity – as long as it is recognised for what it is: a relative assessment of the value of owning something today against owning it at some point in the future.
Andy is a highly experienced industry professional and market risk management expert with a background in oil refining. He is a trained Chemical Engineer with hands-on experience of the physical demands and logistics of the entire process of turning crude oil into an end product.