South Korean regulators are turning to artificial intelligence to help detect illegal use of algorithmic trading in financial markets.
The Korea Exchange has temporarily hired outside experts to develop a model to monitor algo trades, according to Lee Jong-pil, an official at the bourse. They are expected to complete studying the model by October, and the exchange will then test and decide whether to use it, he added.
“We intend to crack down on bogus orders or stock manipulation with this new model,” Lee said by phone on Tuesday. “With algorithmic trading widely used in markets, unfair trades using algos are allegedly pervasive.”
Authorities aim to stamp out specific trading patterns, such as market-making, spoofing or layering with the help of the model, according tostatement on KRX’s website in April. Spoofing is a practice where traders enter orders they don’t intend to execute tomislead others in the market as to the demand to buy or sell. One type of spoofing is called layering, where orders are stacked at various prices on one side of the market.
READ: High-Speed Traders Abandon Microwaves in Korea on Exchange Snub
South Korea’s stock exchange in July 2019 fined Bank of America Corp.’s brokerage arm 175 million won ($148,000) for violating market rules. The U.S. firm disrupted fair market orders through algorithmic trading, helping a hedge fund client earn 220 billion won of profit by repeatedly submitting and canceling quotes for hundreds of stocks in Korea between October 2017 and May 2018, the KRX said back then, describing them as “almostphantom orders.”
“South Korea’s stock market is becoming a good target for algo-trades seeking an alpha recently because many of the traders are retail investors, not professional investors, and they are providing abundant liquidity,” said Bongju Kang, a quant analyst at Meritz Securities in Seoul. “I support KRX’s efforts to monitor algorithmic trading, but I’m not sure how effective it will be. It’s not easy to detect some techniques, such as the one we saw in the BofA case.”
Source: Read Full Article