Study Says Sarao May Not Have Been Responsible for Flash CrashBy and
Trader's extradition hearing takes place in London next week
Stale prices correlate with the onset of the flash crash
Navinder Singh Sarao, dubbed the Hound of Hounslow by newspapers after his arrest for allegedly manipulating markets, has a few academics on his side as he goes back to court next week.
Sarao may not have had a material, or even any, impact on the bout of equity market volatility in May 2010 that later became known as the flash crash, according to a draft research report by University of California, Santa Cruz and Stanford University professors dated Jan. 25. The study, which has yet to be formally released because the authors are still soliciting feedback, claims to be the first to analyze the entire order book on a millisecond level.
The researchers didn’t say whether the 37-year-old Briton is guilty of manipulating Standard & Poor’s 500 Index futures, which is the central plank of the U.S. authorities’ case. And while the authorities have said Sarao contributed to the flash crash, they never identified him as its sole cause.
Sarao could still go to jail for a long time even if he is fully absolved of causing the manic trading episode -- he is alleged to have manipulated markets over several years. He faces a 380-year jail sentence if he is convicted on all counts.
Still, industry experts have questioned how a single trader could have had such an impact on U.S. markets. That the causes of the Dow Jones Industrial Average’s nearly 1,000-point plunge in May 2010 are still being debated reflects dissatisfaction with regulators’ ability to analyze electronic, computer-traded markets that have been described as too complex and as unfair to ordinary investors.
“It’s very important for society to get the diagnosis correct here; if we misunderstand what caused the crash, we can’t do a good job preventing the next one,” said Joseph Grundfest, an author of the report and a former commissioner at the Securities and Exchange Commission.
“The paper was circulated very quietly among a very small number of people. There was no intent or desire to have it public at this stage,” said Grundfest, who is also a professor at Stanford Law School.
Whether Sarao contributed to, or caused, the flash crash matters because it could encourage the government to focus on prosecuting trading strategies such as spoofing, Grundfest, Eric Aldrich and Gregory Laughlin said in the report. Spoofers trick other traders into buying or selling by entering their own buy or sell orders with no intention of filling them.
“Our analysis suggests that this view incorrectly conflates correlation with causation: just because Sarao’s trading occurred at or around the time of the flash crash, does not establish that it helped cause the flash crash,” they said.
The professors also argue that Sarao couldn’t have known in advance that his trading could destabilize markets. That matters because U.S. sentencing guidelines take into consideration whether harm was foreseeable, they said in the paper.
The draft’s release is timely. Sarao appears in a London court next week for a two-day extradition hearing, nearly a year after his arrest. His lawyers are expected to argue that his actions weren’t crimes in the U.K. and, as a British citizen, any trial should take place in the country. Whatever the outcome, the losing side will probably appeal.
Sarao’s lawyers didn’t respond to attempts to contact them about the report.
Sarao’s behavior came to light in April 2015, when he was arrested at his home in the London borough of Hounslow. U.S. prosecutors say that on the day of the flash crash he placed orders amounting to about $200 million that the market would fall, a trade that represented between 20 percent and 29 percent of all sell orders at the time. The orders were then replaced or modified 19,000 times before being canceled in the afternoon.
U.S. authorities have already turned to academia to support their case against the trader. Terrence Hendershott, a University of California, Berkeley, professor pored over almost 400 days of trading data. His research documented Sarao’s trading activity, but didn’t necessarily connect the trader to the flash crash.
The latest research points out that previously “unobserved anomalies” in market-data feeds correlate with the flash crash. The analysis, which is still ongoing, focuses on off-exchange data compiled by the Financial Industry Regulatory Authority’s Trade Reporting Facility.
Millisecond by millisecond, the worst part of the flash crash coincides with the Finra data feed displaying misattributed, late reported trades, the report said. The market began to recover after Finra halted its price feed.
The Finra feed’s oscillations between stale and actual share prices may have been merely symptomatic of the flash crash, but it is possible that the confusion they caused prompted some algorithmic traders to pull out of the market. If the Finra feed was a contributing cause of the flash crash, regulators should “pay attention to data feeds.”
“During periods of high-volume market stress as occurred on May 6, 2010, it is not uncommon for processing queues to develop that create latency in the dissemination of market information, even if trades are reported in compliance with applicable Finra rules, as we believe was the case during the time period examined by the paper,” Finra said in a statement Wednesday.
“With such a complicated, swift and high-message traffic event that occurred on May 6, it is a challenge to discern correlation from causation,” the regulator added. “Accordingly, we are pleased to continue to work with Mr. Grundfest and his colleagues to offer what insights we can as they finalize their draft research.”
The findings of the research paper were largely in line with a joint report from the SEC and the Commodity Futures Trading Commission. That publication said sell orders by a mutual fund exacerbated an already wobbly market.
To continue reading this article you must be a Bloomberg Professional Service Subscriber.
If you believe that you may have received this message in error please let us know.