Commodity trading houses are increasingly juggling new technologies and large amounts of data — which has itself become a commodity — but ultimately it is still the trading desk personnel's responsibility to check final documents such as bills of lading to ensure security and eliminate fraud, participants in the Financial Times' Commodities Global Summit said Sept. 28.
Trading still relies heavily on physical documents, blockchain does not give complete security, mistrust has led some to move away from algorithmic trading and some banks — including BNP Paribas last week — have preferred to pull out of commodity financing rather than push for technology adoption, executives from Vitol, Mercuria and The Metals Risk Team said, highlighting the challenges that new technologies have brought.
Turning data into information
"These days, we're swamped with data and the name of the game is how do you turn this massive amount of data into information? ... Some people say data is the oil or the energy of the future," Vitol Chief Information Officer Gerard Delsad said during a panel titled Commodity Trading in the Digital Era.
Over the last three years, Vitol has been building "a big data refinery which collects raw data, crosses it, cleans it and adds some proprietary data — which is less every day — storing it and making it available through an [application programming interface] for machine learning, for the business, the front office, then [artificial intelligence] becomes the cherry on the cake ... then you can really deliver some value ... But there's a lot of work in organizing it, cleaning it and making sense of it ... it's very qualitative ... if you're able to make a decision faster that is useful," Delsad said.
Emphasis growing on demand, not supply data
Machine learning has been particularly useful due to the changes brought by the coronavirus pandemic. "In the past, the focus has been much more on understanding supply, and nowadays with COVID, the emphasis is increasingly on understanding demand ... the timing of the pick up by location and by product in all parts of the world ... in this area analytics and predictive tools have really shown their value," Delsad said.
To deal with the onslaught of data, trading houses are now employing ever-more data scientists in assisting the trading desks with tools – with around 30 nowadays at Mercuria and more than 40 at Vitol, executives said. Data involves "a trade-off between quality and quantity," said Cyril Reol, chief information officer at trader Mercuria.
A number of commodity trader bankruptcies due to fraud have been uncovered in Singapore and elsewhere in Asia in recent months, with duplication of bills of lading from different banks, FT commodities editor Neil Hume said.
Technology could help stamp out some fraud in proof-stamping, and this could involve blockchain or other distributed-ledger technologies, said Reol, whose employer is a shareholder in some blockchain platforms, adding that traders are working together to help major oil and other companies "get digitalized" and cut out the middleman.
"Adoption will come, but this will take time," Reol said, noting that even with the use of blockchain systems, personnel were still need to review all documents.
Delsad said blockchain could be useful for reimaging documents, but he does not necessarily see added value in its use if traders already have an information system in-house.
Robert Fig, partner in The Metals Risk Team, a metals price risk management company, said that while blockchain is part of frequent discussions at the broader level "in itself (it) doesn't resolve the issue of security, which is an area we seriously need to look at ... as frauds manage to get through."
However, the vast amounts of paperwork involved in commodity trading could be simplified via the use of new technology, he said.
Brokers 'squeezed' out of market
Fig noted that there has been a "pretty dramatic" reduction over the last four to five years in the numbers of banks and brokers in the commodities markets, partly due to the capital adequacy rules that have squeezed many brokers out of the market and created a move away from banks to traders. "The fact there are fewer brokers has made it very difficult for those using the market [for hedging] to manage their risk sufficiently ... ironically the capital adequacy rules of Basel have in fact created more risks than they solved in the market."
Hedging has become very computerized, using algorithmic trading, which has squeezed some traders out of a market seeking greater transparency. However, on some markets such as the London Metal Exchange, algorithmic traders are not seen in the same way as in other futures markets and some people are moving away from electronic trading and moving back to the telephone market, where they're forced to go back to technologies used in years past in a search for greater liquidity, Fig said.
Mercuria and Vitol may be considered huge hedgers, but are not currently using algorithmic trading, and machines are far from taking over from human beings at the big trading companies, the executives said.
