One in a Million: SEC Data Analytics and Cherry-Picking
After the simulation was run one million times, the SEC knew the advisory firm owner could not have profited that much by chance.
The Commission on June 29 charged Wisconsin-based adviser Welhouse & Associates and its owner, Mark Welhouse, with cherry-picking. Specifically, according to the administrative order instituting proceedings, the two improperly allocated to Mark Welhouse’s personal and business accounts options trades that appreciated in value during the course of a trading day, while allocating to clients trades that depreciated in value.
The enforcement action is the first under a data-driven initiative to identify potential cherry-picking – fraudulent trade allocations under which the adviser keeps the best trades for itself, the SEC claimed. Under the initiative, enforcement investigators from the SEC’s Division of Enforcement work with economists from the agency’s Division of Economic and Risk Analysis to "analyze large volumes of investment advisers’ trade allocation data and identify instances where it appears an adviser is disproportionately allocating profitable trades to favored accounts," the agency said.
"Cherry-picking schemes can be extremely difficult to detect without an investor astutely noticing that something may be amiss and coming to us with a complaint about the adviser," said Enforcement Division Asset Management Unit co-chief Julie Riewe. "We devised this initiative to identify specific custodians providing services to investment advisers and their clients and leverage their trading records and other data to efficiently target preferential trade allocations occurring outside the detection of even the most observant client."
Welhouse and his firm were accused by the SEC of receiving more than $442,000 in ill-gotten gains from February 2010 to January 2013 by unfairly allocating options trades in an S&P 500 exchange-traded fund. According to the administrative order, Mark Welhouse’s personal trades had an average first-day positive return of 6.28 percent, while his clients’ trades in the same option had an average first-day loss of negative 5.05 percent.
Just to be sure that Welhouse and his firm could not have received such profits "from a coincidence or a lucky combination of trades," the SEC staff conducted a statistical analysis involving a simulation, which it ran one million times. The result was that the profits could not have resulted from luck, the agency said. Mark Welhouse’s profit "was substantially higher than every one of the one million random simulations. The results show that there is only an infinitesimal likelihood of achieving by chance a profit like Mr. Welhouse’s."
Further, the SEC said, when comparing the proportion of profitable trades allocated to Mark Welhouse’s accounts to the proportion of profitable trades allocated to his clients’ account, the likelihood that Mark Welhouse’s personal account would receive such a high proportion of profitable trades by pure random chance "is less than one in one trillion."
The role of data analysis
That said, did the simulations by themselves lead to the charges? Not according to Zaccaro Morgan partner Nicolas Morgan, who attributed the SEC’s enforcement action more to the termination of the Welhouse account by its custodian broker following several warnings from the broker to Welhouse (see below). "It appears that the SEC’s statistical analysis was done after the fact to determine that the transactions’ profitability was ‘statistically significant’ and that the odds of achieving the same profits by chance were ‘infinitesimal,’" Morgan said. "This is the sort of statement an expert witness would make in an effort to show a fraudulent intent. However, merely eliminating chance as an explanation for trade allocations does not prove fraudulent intent."
"The case really comes down to a difference between the adviser’s representations to investors and actual practice," he said," noting SEC disclosure allegations involving the adviser’s Form ADV and firm brochures. The SEC doesn’t need statistics to prove any of that."
Stern Tannenbaum partner Aegis Frumento also dismissed the idea that the SEC’s data analysis led to the cherry-picking discovery. "The administrative order is quite clear that the preferential allocations were so obvious that the broker/custodian called the adviser on it multiple times, and then terminated the relationship when the adviser wouldn’t stop doing it," he said. "The only ‘data-based’ exercise performed here was an after-the-fact simulation that showed the adviser’s profitable trades could not have resulted from random chance – a nice statistical exercise but it only corroborated what the adviser’s brokers already knew and what was fairly obvious just from looking at the difference in returns between the personal accounts and the client accounts. That analysis didn’t detect this conduct, and for the SEC to suggest that it did, as its press release clearly does, is simply misleading."
Mark Welhouse allowed SEC staff to interview him in January 2014, and that the interview could be recorded, according to the administrative order. During the interview, he said that he used one master account for trades in his four personal accounts and used a different master account for his clients’ trades, the agency said. But the agency also reported him as saying that "there were times when he allocated … options trades from the client master account to his personal accounts," something he labeled as "mistakes." He also acknowledged that his broker called him many times and expressed concern about the allocations, the SEC said.
The SEC also claimed that, in his recorded interview, Welhouse stated that all his trades were allocated on a pro-rata basis, and that this was made clear in his January 2012 Form ADV.
But, the agency charged, "contrary to Welhouse’s policies and procedures and its Form ADV statements, Mr. Welhouse, on behalf of Welhouse, did not allocate … options trades pro rata." Instead, he allocated a disproportionate number of profitable options trades to accounts belonging to himself or another person sharing his last name, the SEC said, by purchasing the options in a master account at his broker and delaying allocation of the purchases until later in the day, after he saw if the securities appreciated in value.
"During the relevant time period, approximately 58 percent of … options trades occurred before 11:00 a.m., while about 58 percent of … options trades were allocated to accounts after 2:00 p.m.," the SEC said. "Moreover, approximately 47 percent of … options trades were allocated to accounts after 3:00 p.m., during the last hour of regular market hours for options trading. This delay allowed Mr. Welhouse to selectively allocate profitable trades to his personal accounts.
The broker notices
The SEC staff, which apparently reviewed the broker’s internal compliance notes, listed the following:
In April 2010, an employee of the broker told Mark Welhouse that the broker was monitoring his trade allocations. "During this conversation, Mr. Welhouse agreed to separate his personal and client trading in different accounts," the SEC said.
Following the April 2010 conversation, the broker’s trade allocation surveillance system "flagged Mr. Welhouse’s joint account nine times between May 2011 and September 2012."
In February 2012, another employee of the broker called Mark Welhouse because he "seemed to be making preferential trade allocations from his clients’ master account to his personal account," the SEC said. "Mr. Welhouse returned the employee’s call, and, during the recorded telephone call, the employee reminded Mr. Welhouse to keep his personal trading separate from his clients’ master account, and Mr. Welhouse agreed he would do so."
In June 2012, another brokerage employee called Mark Welhouse and told him that he was continuing to allocate trades to his personal account from his clients’ master account, which had the appearance of preferential trade allocation. "The employee reminded Mr. Welhouse of the two prior conversations on the same issue, and the employee told Mr. Welhouse that the broker would consider blocking allocations from a master account to his personal accounts if the practice continued," the agency said.
In September 2012, the broker flagged Mark Welhouse’s trade allocation a ninth and final time.
In December 2012, the broker terminated its relationship with Mark Welhouse.
The SEC staff interviewed three Welhouse clients who experienced significant investment losses on options trades, including unprofitable first day returns. What the staff learned, according to the administrative order, was that Mark Welhouse’s clients were not aware of his trade allocations, or that he was even trading options in their accounts. Welhouse clients were typically individuals and families, the agency said, and each of the three interviewed "considered himself or herself to be an inexperienced investor seeking a conservative approach in managing his or her accounts."
According to the administrative order, when one client asked Mark Welhouse about the losses, "Mr. Welhouse told the client that he had experienced the same losses in his personal accounts."
"Advisers should keep their personal trades separate from their clients – in separate accounts," said Frumento. "They can make different trades in their own accounts than they make in their client accounts, but doing so always risks being questioned whether their personal trades are at the expense of their clients. Better to adopt the same strategies for both, trade the same way in parallel. Commingling personal and client trades in one master account and then allocating is always a dangerous practice."
What’s next and violations
The case now moves to an administrative hearing, which must be held by approximately the end of August.
The SEC charged Welhouse and Mark Welhouse with willfully violating Section 10(b) of the Exchange Act and its Rule 10b-5, which prohibit fraud. In addition, the two parties were charged with violating Sections 206(1) and (2) of the Advisers Act, for allegedly committing fraud and for allegedly making false or misleading statements on the firm’s Form ADV. No attorney was listed by the SEC as representing the firm, nor could one be found during a web search. A call to the firm itself reached a recording saying the phone number was no longer in service.