As for the need for an algo to trade spreads, spreads are traded for the same reason today that they were 40 years ago. Only a very small percentage of established flat positions are maintained into the delivery period, the perpetual longs and shorts either exit their positions or roll their positions into a future month. All market conditions considered, it is most likely that that the participants with perpetual positions will roll their positions forward, hence the sensitivity to the spread. In general, the best-case scenario for the long is to roll their position at something close to even money or an inverse. Think of a long establishing a position in crude oil with an average price of $50. Every time that the long is able to roll at an inverse of $1, the average price of their long is reduced by $1. Over the course of a year, the entry price is reduced by $12 which is very significant. The opposite applies to the perpetual short, they want to roll their position at a value closer to full carry. Although it is very easy to understand why spreads exist, it has been my experience that too many participants very it as little more than a necessary evil, participants use time as the only parameter to roll their positions, often at a very significant cost. I can recall numerous spreads that were mismanaged, in the case of 2020 crude the cost was probably in the hundreds of millions. No excuse for something like that to happen.
Calendar spreads are sensitive to those fundamentals that affect the delivery process. I have used this approach to establish positions that have represented 10-50% of the nearby open interest (offset by an equal and opposite deferred position). The strategy that I have used has been taken from observing similar positions being taken by other participants, albeit with modifications. Perhaps the contracts that this strategy is most applicable to are the grain contracts. Given that North American production is concentrated in a relatively small window, there is inherently an advantage to estimating grain stocks that are in a deliverable position as compared to another commodity such as crude. Drawing carryout from a balance sheet gives an estimate of those supplies that may be delivered against the futures. Hence the annual changes from large inverses to markets displaying carrying charges from one year to the next (carryout used as a proxy for deliverable supplies). A large part of the strategy but not to be used in isolation.
A current example of carryout influencing spreads is being played out in the corn market. Whereas several months ago the market was perceived as one with burdensome supplies and characterized by spreads at significant carrying charges, strong exports have decreased the carryout and reduced the threat of delivery. As a result, the old crop/new crop spread closed at a small inverse after being at a large carry during harvest. The maturation of this spread will depend on continued end user support balanced against the threat of tariffs (an example of how government policy can influence a spread). For those perpetual shorts that use time to roll their positions, the market presented them with an opportunity to roll their short during the fall period, now they are faced with buying the spread at an inverse. From my perspective, this is where an automated system using some of the criteria I have outlined would be of significant value to a hedger or fund. I don't have the technical expertise.
Further to the old crop/new crop corn spread, the USDA reduced carryout to 1.738 b bus, a decrease to 200 m bus. The market responded by the spread inverting another nickel. In terms of how I have framed spreads, there are less stocks available to be delivered against futures, advantage goes to the perpetual long. From a flat price perspective, being short new crop at current levels (close to cost of production) does not make a lot of sense. However, the upside is also limited by the perpetual short who will be buying the old crop/new crop spread. It now appears that the large carries that the market experienced a couple short months ago will not be retraced this crop year. Once again, for those hedgers with perpetual positions, allowing time to dictate when positions are rolled is a failing risk management strategy. However, some things never change and when it comes to trading futures it is a characteristic of the market that has been consistent over the past 40 years.
In the first quarter of 2022 Cenovus, a Canadian oil company, lost 970 million that it attributed to hedging activities. Furthermore, it stated that it would suspend further hedging activities with particular emphasis on WTI. How does a hedger allow losses to reach those levels? Cenovus could be classed as a perpetual short (that is only a term that I have used, someone else might call it something else) that is, long cash - short futures. Obviously they had significant exposure when the Russia/Ukranian war broke out if I recall correctly the April/dec WTI spread inverted to 15-20$. For those perpetual shorts positioned in the nearby, they had the unenviable task of rolling their short positions at large inverses at a time of great uncertainty. In theory their cash position should have offset their losses in the futures, but obviously that didn't happen. The likely reason is that Canadian oil is of a different quality and trades at a discount to the WTI, what is referred to as the "basis". When the war started the basis must have weakened relative to the futures, thus not providing for a "perfect" hedge. The losses by Cenovus (and other energy companies) relate to an inadequate trading strategy from at least a couple of approaches. Firstly, it appears that little consideration was given to their excessive exposure to the nearby at a time when a major war was looming. In the least consideration should have been given to rolling to a deferred position rather than waiting for a margin call. Secondly. the WTI provided a less than perfect hedge, i.e. the cash product not highly correlated with the futures, at a time of increased volatility the discrepancy becomes painfully obvious. This is a common dilemma faced by foreign entities that are perpetual participants, all is good until it isn't. Another example for an automated system to trade spreads, this example for hedgers.
The approach I have described above was developed during the 20 years I worked on a trading floor as a local. All commodity futures contracts require a delicate balance between commercial and spec participation. During the open outcry era, locals provided a significant percentage of the speculative positions via a variety of strategies. Ultimately the job of the local was to assume the risk that the commercial participant was attempting to reduce. For myself, I found that I could manage risk most effectively using spreads. That is, I could assume the flat price risk hedgers were looking to reduce by offsetting a particular position with an equal offsetting deferred position. Some of the criteria I have described above that ultimately affects the ability to take/make delivery determined entry and exit. In the end, while I was able to assume the hedger's risk, I was rewarded by the perpetual long/short rolling their positions forward. Going back to the open outcry era, it is safe to say that there was one or more individuals on each trading floor that used a similar strategy. I can only speak for myself but I recorded 10-12 years of ROI ranging from 100% to 1000%. Moving forward, I am open as to how to move this strategy forward,