Springer Virtual Barrels: Quantitative Trading in the Oil Market

Best book on oil trading. Oil 101 is always touted as the bible, but this is more relevant.

A thorough, insightful and digestible glimpse into the oil complex and systematic approaches to trading it.

“Virtual Barrels: Quantitative Trading in the Oil Market” by Ilia Bouchouev is a fantastic book, which I greatly enjoyed reading. For me, it had the perfect mix of digestible mathematical exposition and practical industry insights. I also appreciated his reference to older economic and mathematical thinking on financial derivatives by Keynes and Bachelier which other treatments of the topic probably would have skipped. While the book focuses on the oil industry, some of the lessons might – suitably modified – also apply for other commodities. Forward prices of commodities In the textbook treatment of forward commodity pricing, the following argument is typically made. A hedger (e.g., a refinery purchasing crude oil, or a pizza manufacturer purchasing cheese) wants to enter an agreement today to purchase a certain amount of a commodity (crude oil, mozzarella cheese) in, say, 3 months’ time at a known price, F. A trader or bank offering to sell the commodity at price F would then, to ensure it can deliver on the agreement, borrow the amount of money needed to purchase the commodity in the spot market at price S and store it for three months. With the borrowed money, the trader then buys the commodity in the spot market and stores it a warehouse at up front cost C. Hence, at the time of delivery the bank can take the commodity out of storage, deliver it to the hedger and obtain the agreed upon forward price F. At the same time, the trader will need to pay back the money it borrowed, which is equal to (S+C)×(1+r(3/12)), where r is the three month annualised interest rate. If the forward price exceeded this amount, competitive traders would try to undercut each other competing for hedgers’ demand. Likewise, if the forward price were below this amount no trader would be willing to offer it. In short, the equilibrium forward price is given by F = (S+C)×(1+r(3/12)). This equation (in its generalised form) implies that the forward price is above the spot price. While this is generally true for stocks and bonds, it does not hold for commodities like oil and dairy products. Storage and the Behavior of Commodity Prices Bouchouev follows the economic literature on commodity markets and explains the presence of backwardation by the impossibility, at the aggregate level, to have negative inventory of commodities. With a positive level of inventories, when the spot price is above the forward price, a trader could “borrow” the commodity, sell it in the spot market, invest the proceeds at the risk-free rate of interest, purchase the commodity forward and then, at the delivery date take the forward purchased commodity to return the borrowed commodity. But when there are no inventories, there is no stored amount of the commodity that could be borrowed and sold in the spot market. Hence, the arbitrage relationship breaks down, which may result in backwardation. In addition, for some commodities there may also be an upper bound on the level of storage that may lead spot prices to lie far below forward prices. While some commodities may be easily stored at little extra cost, crude oil cannot just be left lying around somewhere. Hence, when storage facilities are at close to their maximum capacity, current production cannot be transferred into the future but must instead be used today. This (or the expectation of storage capacity running out shortly) may then even lead to prices dropping below zero, something that has occurred in oil markets in the wake of a drop in oil demand due to the covid epidemic (which Bouchouev discusses in more detail) and something that regularly occurs in electricity markets with a large share of non-interruptible renewable supply (e.g. in Germany). Hedging Demand Separate from considerations of storage, the different needs of buyers and sellers of a commodity also matter for the structure of the forward curve. Typically the seller of a commodity like oil or cheese is more exposed to fluctuations in its price than a buyer, since the latter typically has many more inputs it purchases. As such more sellers, whose natural forward position is long, are likely to participate in financial markets. To offset their natural forward position, sellers will take short positions in forward markets which due to the lower participation of buyers will tend to push down forward prices compared to a neutral position. Hence, the differential hedging pressure may create a downward bias in forward prices, leading to a situation of backwardation. Bouchouev explains that this was initially the case in financial markets but that the resulting returns for traders from buying oil forward while selling it closer to the spot market reduced the extent of backwardation over time. In addition, the correlation of the price of oil with inflation also led other asset managers to want to gain exposure to oil prices for general asset management reasons. Cointegration and Stat Arb One of the first lessons in time series analysis is that regressing one variable with a trend on another can lead to spurious statistical significance. To test whether the relationship is meaningful one can take the estimated relationship and calculate the deviation of each observation. If these residuals themselves are non-stationary (i.e. do not have a trend) then this suggests that the relationship between the original variables is meaningful. I have found cointegration analysis to be a powerful (and very cool) tool, not just in analysing financial markets and trading strategies, but also in assessing relevant markets for regulatory and antitrust work. Bouchouev’s book is about trading strategies, not formal econometric analysis, so unfortunately he does not go into too much mathematical detail. He mentions, however, some possible cases of cointegrated time series: 1) the WTI (West Texas Intermediate) and Brent oil price benchmarks, 2) crude oil and the value of the refined products, 3) different regional indices. Because of changing fundamental regimes these cointegration relationships may be subject to structural breaks. The most striking example Bouchouev offers is how from around 1993 to 2008 the WTI and Brent prices moved closely together. This was a period in which the US imported oil, so its domestic prices were determined by world prices. From 2008 until the end of 2015, the shale oil boom in the US drove down US domestic prices (and thus also the WTI price). Since oil exports were prohibited until December 2015 there was also no economic pressure pushing both price benchmarks closer. Options The aspect that I perhaps liked most about Bouchouev’s book is his explanation for the demand of oil industry participants for financial options. For example, shale oil producers are able to flexibly expand production if the price of crude oil exceeds their production cost – in other words a shale oil producer effectively holds a call option on the price of crude oil with the strike price given by its cost of production. Shale oil producers might then sell a corresponding financial option to offset their natural position and thereby obtain a premium payment that can be used to complement more traditional methods of finance.

I work in the financial industry and this book is really excellent. It strikes a great balance between mathematical and financial information. The explanations of the dynamics of the Oil futures curve and how to think about markets is very solid. There is also a great overview of various signals typically employed in the market. I thoroughly recommend this book to anyone interested in trading oil and in understanding how to think about commodity markets more generally

Definitely recommended for a number guy when trying to understand how to trade commodities!