BY CONTINUING TO USE THIS SITE, YOU ARE AGREEING TO OUR USE OF COOKIES. REVIEW OUR COOKIE NOTICE
X


Commodity market AI applications are emerging along with new risks

New York (Platts)--8 Nov 2018 547 pm EST/2247 GMT


AI can streamline trading and help regulators

Smart contract opportunities tempered by challenges

Artificial intelligence and smart contract technology like blockchain are slowly being adopted by commodity markets, creating opportunities to streamline trading and other functions, but not without introducing challenges and risks experts said Thursday.

"Price suggestion has clearly become a key factor [for AI] where for large trades and complex derivatives it used to take a while to price trades," Sandeep Arora, chief operating officer of Citi Markets and Securities Services at global bank Citi, said at the International Swaps and Derivatives Association Technology Forum 2018 in New York.

"Now we are making that information much more quantitatively available to our traders to make faster and better decisions" with the help of AI, Arora said.

And while AI could theoretically be used to forecast power, gas or oil prices, some market participants are using the technology to better inform theories used by humans to forecast prices.

"We are not so much interested in predicting prices, we are interested in developing better theories that can help us make better predictions of prices," Marcos Lopez de Prado, principal and head of machine learning at AQR Capital Management, said.

An asset manager can develop better theories and better identify risk using these technologies, he added. For example, broad market narratives can be identified by scanning millions of daily news articles to see what people are reading and what they are worried about, he said.

You can develop a delta between a decision made by a human and a machine based on the machine's ability to scan a larger sample of market news, Lopez de Prado said.

MARKET MANIPULATION

While keeping up with technology can be a challenge for regulators, AI could also be used to spot patterns of market manipulation. For a human to identify whether an order entered into an exchange has the intention of disrupting or manipulating a market one would need to look at millions of examples and terabytes of data to find patterns or trends of manipulation, Lopez de Prado said.

Algorithms can uncover the footprint a human leaves when manipulating markets. So a machine can sift through huge volumes of data to pick out questionable trades that can then be more deeply scrutinized by a human, he said.

Legal experts described ways in which traditional paper contracts can be converted into smart contracts based on distributed ledger technology like block chain.

When you look at a trade, there is the intended outcome and all the things that could go wrong like insolvency or non-payment and contracts typically deal with those potentialities, Harry Jho, partner at securities finance law firm, Harry Jho, said.

While ambiguous language that cannot be written by computers is needed for the unforeseen circumstances that could arise, distributed ledger technology could be used for the standard portion of the contract, he said.

One challenge is expressing the coded portion in a way that could be read and understood by a judge, Paul Lewis, partner at law firm Linklaters, said.

Another challenge associated with using smart contracts in commodity market trading could involve a situation where a trade in a smart contract is supposed to hedge something that is not in a smart contract, the panel's moderator Scott Farrell, partner at law firm King & Wood Mallesons, said.

Dickson Chin, partner at Jones Day raised another potential risk associated with the growth of smart contracts, which is the distributed ledger platform itself. It appears we will likely have permissioned platforms for derivatives and smart contracts in the future, but this introduces a third element to the contractual relationship, which is the platform itself.

He wonders what will happen if the underlying platform becomes insolvent. -- Jared Anderson, jared.anderson@spglobal.com

-- Edited by Rocco Canonica, newsdesk@spglobal.com




Copyright © 2018 S&P Global Platts, a division of S&P Global. All rights reserved.