What is the information content of oil futures curves?

Julian Reynolds

Moves in oil prices have significant implications for the global economic outlook, affecting consumer prices, firm costs and country export revenues. But oil futures contracts tend to give an imperfect steer for the future path of oil prices because, at any given time, futures contracts may be affected by a wide range of fundamental drivers, besides the expected path of future spot prices. This post presents an empirical methodology to determine the so-called ‘information content’ of oil futures curves. I decompose the oil future-to-spot price ratio into structural shocks, which reflect different fundamental drivers of futures prices, in order to identify the extent to which futures prices reflect market information about the outlook for spot prices.

What are the fundamental drivers of futures prices?

A futures contract is an agreement to buy or sell a given commodity at a given point in the future, at a predetermined price. In general, futures prices are driven by market expectations about future market conditions, as well as speculative activity by investors. Oil is a particularly heavily traded commodity: as a physical asset, spot prices reflect current and expected future supply and demand; and there is a deep market for oil futures.

The expected future spot price of oil is invariably a key determinant of futures prices. A number of forecasters thus use futures prices as a signal of the likely path of future spot prices. But there are several other fundamental drives of futures prices that may be distorting this signal, as summarised in Table A.

Table A: Drivers of futures prices

Source: Nixon and Smith (2012).

All else being equal, the futures curve becomes more upward sloping as risk-free interest rates rise, because the risk-free rate is the opportunity cost of holding the futures contract.

In the other direction, risk premia is expected to weigh down on oil futures prices. Oil is a risky asset, which means that expected returns reflect a (typically positive) risk premium. This risk premia will put downward pressure on observed futures prices relative to the unobserved expected future spot price, as investors will only pay below expectations of future prices to compensate for the risk that prices fall. This makes the futures curve downward sloping, consistent with Keynes’ ‘normal backwardation’ hypothesis (Till (2006)).

Oil is also a physical asset, which means that the ‘convenience yield’ and storage costs also affect futures prices. The convenience yield is the benefit accrued only to holders of physical commodities, who can smooth through demand shocks by boosting supply at short notice. The higher the convenience yield, the easier it is for commodity holders to smooth through shocks. This disincentivises holding the futures contract relative to physical commodities, weighing on futures prices. Set against this, holding physical commodities imposes storage costs. A rise in storage costs would be passed on to commodity holders, which increases investors’ incentive to buy futures contracts instead, resulting in higher futures prices.

The convenience yield and storage costs are individually unobservable. But the ‘net convenience yield’ – which equals the convenience yield minus storage costs – can be measured by the ratio of futures prices to spot prices, minus risk-free interest rates.

Finally, higher oil inventories tend to push up futures prices. This is because inventories tend to be negatively correlated with the net convenience yield (Chart 1), as Gorton et al (2007) suggest. Intuitively, at low levels of inventories, commodity holders have less capacity to smooth through shocks by running down stocks before they run out altogether, so they have a stronger incentive to increase holdings of physical commodities relative to futures contracts. In addition, storage costs are lower, because there are fewer commodities that require storage.

Chart 1: Net convenience yield and oil inventories

Note: Net convenience yield equals the two-year oil future-to-spot price ratio (annual average) minus the two-year US Treasury bill rate.

Sources: Bloomberg, Eikon by Refinitiv, International Energy Agency and Bank calculations.

How to identify the drivers of futures prices?

My analysis aims to distinguish whether moves in oil futures prices reflect market expectations about future spot prices or other fundamental drivers.

The variable of interest is the slope of the oil futures curve. I capture it using the ratio of the futures price to the spot price for a given maturity (hence future-spot ratio), expressed as an average annual percentage difference. I estimate a structural vector autogression model, to examine how the future-spot ratio moves with: i) similar maturity US treasury yields, which are a proxy for risk-free rates; ii) oil-implied volatility (OVX) as a proxy for risk premia; and iii) the level of OECD oil inventories. I estimate the model using monthly data from 2003 to 2022, and include a linear time trend.

I use ‘sign restrictions’ to identify structural shocks within the model, as listed in Table B. These shocks represent different fundamental drivers of the future-spot ratio, in accordance with economic theory. Specifically, I identify a structural shock depending on the direction in which I expect certain model variables to comove in response to this shock, during the same month that the shock occurs. Finally, I calibrate impulse response functions, the response over time of the two-year future-spot ratio to the structural shocks, as shown in Chart 2.

Table B: Sign restrictions and structural shocks

Source: Authors’ calculations.

In the first row of Table B, the information shock is associated with higher expected future spot prices. This causes future-spot ratio to increase, and investors build up greater inventories in anticipation of higher prices. I also find there is a positive correlation between OVX and the future-spot ratio in my sample, so the information shock is also associated with a rise in volatility. A one standard deviation (1std) information shock causes a 3 percentage points rise in the future-spot ratio on impact (Chart 2, aqua line).

In the second row, the interest rate shock is associated with a rise in both treasury yields and the future-spot ratio, as higher risk-free rates lead to higher returns to holding a futures contract. A 1std interest rate shock causes a 1.6 percentage points rise in the future-spot ratio at the peak (orange line).

In the third row, the risk premium shock is associated with a fall in in OVX and a rise in the future-spot ratio. This shock is consistent with the theory outlined by Nixon and Smith (2012), whereby reduced risk premia leads to higher future prices. The future-spot ratio rises by 1 percentage point at peak (purple line), four months after the shock materialises.

In the final row, the convenience yield shock is associated with higher inventories, lower treasury yields, and a rise in the future-spot ratio. In other words, greater inventories lead to a fall in the net convenience yield, causing futures prices to rise. This shock causes a 1.1 percentage points rise in the future-spot ratio at peak (gold line).

Chart 2: Impulse response of future/spot ratio

Note: Solid lines denote the median of the sample of impulse responses. Dashed lines denote a one standard deviation confidence interval.

Source: Authors’ calculations.

What is the information content of oil futures curves?

Chart 3 illustrates the historical decomposition of the two-year oil futures-spot ratio into the structural shocks identified using sign restrictions. The decomposition of the one-year and three-year oil futures contracts appears very similar to the two-year contract.

Chart 3: Decomposition of future/spot ratio

Note: Residual denotes the difference between the two-year futures-spot ratio and the four structural shocks identified using sign restrictions.

Source: Authors’ calculations.

The chart shows that information shocks (aqua bars) have been a significant driver of the futures-spot contract for much of the time sample. For instance, during the Global Financial Crisis, beliefs that oil prices would rebound after a sharp slump drove most of the increase in the future-spot ratio. Conversely, the decrease during 2018 was likely driven by beliefs that oil prices would fall. This exercise suggests, therefore, that futures curves generally embed a high degree of information about the outlook for oil prices.

However, there have also been some noteworthy examples where the future-spot ratio reflected changes in fundamentals. From August 2014 to August 2017, when oil inventories were steadily increasing, the upward sloping futures curve was roughly evenly driven by interest rate (orange bars), risk premia (purple bars), convenience yield (gold bars) and information shocks. In addition, convenience yield and risk premia shocks were the main drivers of the downward sloping futures curve in 2021, when inventories fell sharply to an eight-year low.

Sensitivity analysis highlights the uncertainty associated with this exercise. My results are robust to the choice of time trend or lag length. But they appear somewhat sensitive to the specification of sign restrictions and choice of explanatory variables. If I relax the restriction that inventories increase when an information shock materialises, information shocks typically become less important drivers of the future-spot ratio, relative to convenience yield shocks. Conversely, using equity-implied volatility as a proxy for risk premia means that convenience yield shocks become much less prominent. On balance, it is reassuring that my central case results lie between these outcomes.


This post presents an empirical exercise to examine the information embedded within oil futures prices. My results suggest that the slope of oil futures curves generally reflects a high degree of information about the outlook for oil prices, even after accounting for the impact of fundamental drivers. This kind of exercise can be valuable to judge how much weight to place on futures contracts as an indicator of expected future spot prices. However, it remains challenging to accurately forecast oil prices, which will be strongly affected by unforeseen future shocks.

Julian Reynolds works in the Bank’s International Division.

If you want to get in touch, please email us at bankunderground@bankofengland.co.uk or leave a comment below.

Comments will only appear once approved by a moderator, and are only published where a full name is supplied. Bank Underground is a blog for Bank of England staff to share views that challenge – or support – prevailing policy orthodoxies. The views expressed here are those of the authors, and are not necessarily those of the Bank of England, or its policy committees.