The algorithm-trading whistleblower from Singapore has also turned advisor for the Securities and Exchange Board of India (Sebi), offering it unsolicited advice on how to deal with his complaints.
In a letter written in August, the whistle-blower expresses concerns about lack of action on issues raised by him and offers detailed advice through which wrongdoing can be established.
"I refer to my letter dated January 14, which I believe set off the investigation into market manipulation at the NSE (National Stock Exchange) collocation. It has been eight months since I sent the first letter and I am able to notice some activity on the ground but obviously Sebi has not been able to reach a conclusion," the whistle-blower said.
| UNSOLICITED ADVICE |
|
These letters have been forwarded by the finance ministry to Sebi. Business Standard had reported a study by the regulator's technical advisory committee on March 19. Sebi is yet to pass any order or direction in the matter.
"No one seems to be taking a view of what actually happened in that period and who was guilty and what should be the remedial action," the letter added.
Acknowledging Sebi's limitations, the trader, who has refused to reveal his identity, said an "ideal situation" would be to simply "crowdsource the data of orders and trades during the said period (after masking the identity of the members and not disclosing the client identity either)."
He cited the example of Netflix in the US, which had given out masked user information in a context.
Stating that he had not adequately explained the methodology to be followed to establish his claims in the first letter, the whistle-blower wrote, "The data analysis would require you to parse at least 200 GB of database for each of the dates. Hence, it would be advisable to either have a database expert or Big Data expert for the analysis." Big Data are sets of information that are too large or too complex to handle, analyse, or use with standard methods.
The letter went on to elaborate "the logical way of how Sebi can go about verifying" the allegations.
The Sebi spokesperson did not respond to an e-mail sent on Monday seeking comments on the usefulness of this advice and whether the regulator had acted upon it.
The first step discussed in this "logical way" was to get order and trade logs from the exchange with millisecond timestamps. "In such a case, you would need to only check all the IOC (immediate or cancel) orders fired and see if there is any pattern of first IOC orders being fired for the most volatile periods consistently from one member." If the data showed any one member cornering 25 per cent or more trades, that would be a red flag.
If the millisecond timestamps were not available, Sebi should ask the exchange for a log of who was connected to which TBT server during the three-year period under scrutiny. "If you do get the log, all you need to check is whether there were some servers to which only a few members were connected as compared to other crowded servers," the whistle-blower wrote. He goes on to discuss other options for Sebi, such as Instance Numbers and Fill Ratio.
In the third letter dated October 3, where he had talked about the "dark fibre" links between the NSE and the BSE, the trader shared a discovery about the algo-trading data. "I have come across a significant way to prove or disprove my allegation," he wrote.
"Surprising it came from innocuous research data" provided by the exchange to IGIDR (Indira Gandhi Institute of Development Research).
"The complete data for orders and trades along with timestamp has been provided to IGIDR for the purpose of proving that HFT/algo-trading leads to improvement of liquidity."
This data would be in the order of picoseconds, he estimated.
According to the third letter, IGIDR was provided data with such granularity "far greater than the one required for proving or disproving the allegation."
The simplest way to handle the investigation would be to convert the jiffy timestamp to even say a microsecond one and then see the latency between the various IOC orders generated by the members, the whistleblower said.
