Four recent enforcement actions in the hedge fund space by the US Securities and Exchange Commission (SEC) are the result of warnings generated by a new quantitative model – indicating that a decision to invest in economic analysis is paying off, the agency's chief economist has said.
Four recent enforcement actions in the hedge fund space by the US Securities and Exchange Commission (SEC) are the result of warnings generated by a new quantitative model – indicating that a decision to invest in economic analysis is paying off, the agency’s chief economist has said.
Speaking at the Quant Congress USA conference in New York this morning, Craig Lewis, chief economist and director of the division of risk, strategy and financial innovation at the SEC, described how the revelation of extensive fraud at an investment firm run by Bernie Madoff in December 2008 was the catalyst to developing a model that would flag hedge funds for further investigation.
“We basically said what would be the characteristics that might suggest there is something unusual going on at a particular hedge fund,” said Lewis. “So if you have low volatility and you have steady return performance, the chances are that you are somebody we ought to be taking a deeper look at. They are examples of some of the factors we incorporated into the model - there are a lot of other factors that are more proprietary in nature. But the idea was simply to build a standard regression-type framework and take a panel of returns data and build models around the panel to identify outliers.”
That has resulted in four recent enforcement actions against hedge fund managers - something Lewis described as a key moment for the unit. “This has been one of the watershed moments of the SEC, where people began to see the value of analytics. Historically, the way the SEC has identified and developed enforcement cases has been largely driven by tips, complaints or referrals that market participants submit to the agency - a kind of ad hoc procedure where we rely on the public to identify problems we think are worth additional research. So this is an internally generated model that is trying to do the same thing.”
In developing the hedge fund model, the team looked to create a data warehouse and a report-generating tool that can fire out automated, customised reports to SEC officials. The idea is that these elements can be used for other internally developed models - the division is currently developing an accounting quality tool, for instance.
“What the accounting quality model is designed to do is to look at filers of 10Ks and 10Qs that look a little bit different to everybody else - they may be manipulating earnings. So at its core, it is a tool that detects earnings management among publicly traded companies,” Lewis said.
This kind of analysis isn’t just focused on risk assessment and enforcement – it has also been used to inform Dodd-Frank rule-making. A key driver for this was a decision by a US court to overturn an SEC rule on proxy access last year, on the basis that insufficient cost-benefit analysis had been conducted.
“The agency has decided to try and rectify that situation by improving the quality of the economic analysis that is being used to support our rules,” said Lewis.
As an example, he pointed to a recent rule-making on security swap dealer definitions. By performing an analysis on credit default swap data, and setting some boundaries on what might constitute dealer activity - for example, having a flat book and trading with multiple counterparties - the division was able to put forward significant revisions to the original proposal.
“What we found was that there was a huge break between people who looked like dealers and those who didn’t. We ended up starting out with a threshold in the proposing release that said if you have activity of $100 million, you were going to register as a dealer - and that is notional value. That is approximately 20 trades over a year - it is not a lot of activity. What we found is that if you push the level up to $8 billion, you still capture what we thought was about 97% of all the dealing activity in the market. So it was just simple, fundamental economic analysis using intuition as to how we thought dealers behaved to get to a place where I thought the rule worked a lot better than it did when it was originally proposed,” said Lewis.
This article was first published on Risk