By Kasper Walet, Founder and Partner AI Energizer and Maycroft
I recently read a very interesting article in the Financial Times about how commodity trading firms are still lagging behind other trading sectors in using Artificial Intelligence solutions. but that the industry is trying hard to catch up. These findings are in line with my practical experiences in our advisory services business: AI Energizer.
Particularly in today’s market situation where margins are squeezed, commodity traders are looking for tools that could help them to increase profitability. Intelligent technologies such as robotic process automation and machine learning, offer new opportunities to improve process performance and realize significant cost savings. These technologies can be implemented in short sprints, focused on a specific problem, with manageable costs.
If you click here, you can read our white paper about successful AI projects we executed for two of the leading energy commodity trading firms in Europe.
Compared to other financial and industrial sectors, commodity traders are coming from way behind. Currency, equities and interest-rates investors have already used algorithms, machine learning and artificial intelligence to turn data into successful trades for years. Now, commodity traders are seeking ways of exploiting their information to help them profit from price swings as well. It really is a combination of knowing what to look for and using the right mathematical tools for it.
Traders are looking to gather data on a large scale and run machine-learning algorithms to find patterns linking fundamentals with price movements, thus improve decision-making in trading and, as a result, the profitability. With a properly trained algorithm and a sufficiently sized historical data set, a company using machine learning to identify patterns in the trading data - even when the data has inconsistencies – will reduce redundant trades and streamline the entire process. To accommodate all this commodity trading firms are investing in people, processes and systems to centralize their data.
Despite this new enthusiasm, the road to electronification may not come easily. One issue is that some of the larger commodities traders face internal resistance in centralizing information on one platform. With each desk in a trading house in charge of its profit-and-loss account, data are closely guarded even from colleagues. The move to ‘share all our data with each other’ is a very, very big cultural shift.
Another problem is that in some trading houses, staff operate on multiple technology platforms, with different units using separate systems. Rather than focusing on analytics, some data scientists and engineers are having to focus on harmonizing the platforms before bringing on the data from different parts of the company. Even where the digital infrastructure is in place, it may take some time before AI becomes a large part of commodities trading.
Company leaders should start with the following three steps:
Step 1: Focus on value
As AI can solve targeted problems, it's up to company leaders to identify applications that offer the most potential value. Demonstrating strong returns in a short time will convert the cynics. AI projects typically happen in a series of sprints and can be completed in around 3-4 months.
Step 2: Change Management
Automation alone does not save money or improve performance: People and processes will also have to change. This is an area where there is understandable anxiety; automation stokes fears that machines will take people's jobs. But practice suggests that many of the tasks being automated are activities people tend not to want to undertake, such as spending half the day pulling and loading data. Or they are tasks for which small automation can actually improve people's performance - for instance, by introducing predictive algorithms that help them make better decisions and free up their time for more rewarding, interesting, and higher value-add activities.
This is an entirely new way of working, and company leaders will need to ensure that both they and their people have the right knowledge and skills, such as programming and data science knowledge and process improvement skills - to be successful. And they'll need to ensure that everyone's mind-set and behaviors also shift accordingly.
Step 3: Strategy is King
The backbone of this new way of working is strategy. Companies need to know their own strengths. There will be "keep the lights on" activities that can become touchless, as well as differentiating activities that should become increasingly intelligent. Moreover, companies need to take a fresh look at their organizational structure to ensure that it gives teams the freedom to develop creative solutions and experiment.
Concluding we can say that digitization is increasingly driving trading, and needs to be embraced, as many commodities executives believe. “It’s another tool that traders have to understand.”
Something interesting to share?
Join NrgEdge and create your own NrgBuzz today
The week started off ominously. Qatar, a member of OPEC since 1960, quit the organisation. Its reasoning made logical sense – Qatar produces very little crude, so to have a say in a cartel focused on crude was not in its interests, which lie in LNG – but it hinted at deep-seated tensions in OPEC that could undermine Saudi Arabia’s attempts to corral members. Qatar, under a Saudi-led blockade, was allied with Iran – and Saudi Arabia and Iran were not friends, to say the least. This, and other simmering divisions, coloured the picture as OPEC went into its last meeting for the year in Vienna.
Against all odds, OPEC and its NOPEC allies managed to come to an agreement. After a nervy start to the conference – where it looked like no consensus could be reached – OPEC+ announced that they would cut 1.2 mmb/d of crude oil production beginning January. Split between 800,000 b/d from OPEC members and 400,000 b/d from NOPEC, the supply deal contained a little bit of everything. It was sizable enough to placate the market (market analysts had predicted only a 800,000 b/d cut). It was not country-specific (beyond a casual mention by the Saudi Oil Minister that the Kingdom was aiming for a 500,000 b/d cut), a sly way of building in Iran’s natural decline in crude exports from American sanctions into the deal without having individual member commitments. And since the baseline for the output was October production levels, it represents pre-sanction Iranian volumes, which were 3.3 mmb/d according to OPEC – making the mathematics of the deal simpler.
Crude oil markets rallied in response. Brent climbed by 5%, breaking a long losing streak, as the market reacted to the move. But the deal doesn’t so much as solve the problem as it does kick the can further down the road. A review is scheduled for April; coincidentally (or not), American waivers granted to eight countries on the import of Iranian crude expire in May. By April, it should be clear whether those will continue, allowing OPEC+ to monitor the situation and the direction of Washington’s policy against Iran in a new American political environment post-midterm elections. If the waivers continue, then the deal might stick. If they don’t, then OPEC+ has time to react.
There are caveats as well. OPEC members, who are shouldering the bigger part of the burden, said there would be ‘special considerations’ for its members. Libya and Venezuela - both facing challenging production environments – received official exemptions from the new group-level quota. Nigeria, exempted in the last round, did not. Iran claims to have been given an exemption but OPEC says that Iran had agreed to a ‘symbolic cut’ – a situation of splitting hairs over language that ultimately have the same result. But more important will be adherence. The supply deals of the last 18 months have been unusual in the high adherence by OPEC members. Can it happen again this time? Russia – which is rumoured to be targeting a 228,000 b/d cut – has already said that it would take the country ‘months’ to get its production level down to the requested level. There might be similar inertia in other members of OPEC+. Meanwhile, American crude output is surging and there is a risk to OPEC+ that they will be displaced out of their established markets. For now, OPEC remains powerful enough to sway the market. How long it will remain that way?
Infographic: OPEC+ December Supply Deal
Headline crude prices for the week beginning 10 December 2018 – Brent: US$62/b; WTI: US$52/b
Headlines of the week
The Permian is in desperate need of pipelines. That much is true. There is so much shale liquids sloshing underneath the Permian formation in Texas and New Mexico, that even though it has already upended global crude market and turned the USA into the world’s largest crude producer, there is still so much of it trapped inland, unable to make the 800km journey to the Gulf Coast that would take them to the big wider world.
The stakes are high. Even though the US is poised to reach some 12 mmb/d of crude oil production next year – more than half of that coming from shale oil formations – it could be producing a lot more. This has already caused the Brent-WTI spread to widen to a constant US$10/b since mid-2018 – when the Permian’s pipeline bottlenecks first became critical – from an average of US$4/b prior to that. It is even more dramatic in the Permian itself, where crude is selling at a US$10-16/b discount to Houston WTI, with trends pointing to the spread going as wide as US$20/b soon. Estimates suggest that a record 3,722 wells were drilled in the Permian this year but never opened because the oil could not be brought to market. This is part of the reason why the US active rig count hasn’t increased as much as would have been expected when crude prices were trending towards US$80/b – there’s no point in drilling if you can’t sell.
Assistance is on the way. Between now and 2020, estimates suggest that some 2.6 mmb/d of pipeline capacity across several projects will come onstream, with an additional 1 mmb/d in the planning stages. Add this to the existing 3.1 mmb/d of takeaway capacity (and 300,000 b/d of local refining) and Permian shale oil output currently dammed away by a wall of fixed capacity could double in size when freed to make it to market.
And more pipelines keep getting announced. In the last two weeks, Jupiter Energy Group announced a 90-day open season seeking binding commitments for a planned 1 mmb/d, 1050km long Jupiter Pipeline – which could connect the Permian to all three of Texas’ deepwater ports, Houston, Corpus Christi and Brownsville. Plains All American is launching its 500,000 b/d Sunrise Pipeline, connecting the Permian to Cushing, Oklahoma. Wolf Midstream has also launched an open season, seeking interest for its 120,000 b/d Red Wolf Crude Connector branch, connecting to its existing terminal and infrastructure in Colorado City.
Current estimates suggest that Permian output numbered around 3.5 mmb/d in October. At maximum capacity, that’s still about 100,000 b/d of shale oil trapped inland. As planned pipelines come online over the next two years, that trickle could turn into a flood. Consider this. Even at the current maxing out of Permian infrastructure, the US is already on the cusp on 12 mmb/d crude production. By 2021, it could go as high as 15 mmb/d – crude prices, permitting, of course.
As recently reported in the WSJ; “For years, the companies behind the U.S. oil-and-gas boom, including Noble Energy Inc. and Whiting Petroleum Corp. have promised shareholders they have thousands of prospective wells they can drill profitably even at $40 a barrel. Some have even said they can generate returns on investment of 30%. But most shale drillers haven’t made much, if any, money at those prices. From 2012 to 2017, the 30 biggest shale producers lost more than $50 billion. Last year, when oil prices averaged about $50 a barrel, the group as a whole was barely in the black, with profits of about $1.7 billion, or roughly 1.3% of revenue, according to FactSet.”
The immense growth experienced in the Permian has consequences for the entire oil supply chain, from refining balances – shale oil is more suitable for lighter ends like gasoline, but the world is heading for a gasoline glut and is more interested in cracking gasoil for the IMO’s strict marine fuels sulphur levels coming up in 2020 – to geopolitics, by diminishing OPEC’s power and particularly Saudi Arabia’s role as a swing producer. For now, the walls keeping a Permian flood in are still standing. In two years, they won’t, with new pipeline infrastructure in place. And so the oil world has two years to prepare for the coming tsunami, but only if crude prices stay on course.
Recent Announced Permian Pipeline Projects