The Financial Industry Regulatory Authority (FINRA) just announced it fined Goldman Sachs Execution & Clearing, a division of Goldman Sachs, $800,000 for failing to have policies and procedures to ensure certain clients received the best possible prices when trading in its SIGMA-X dark pool.
FINRA specifically noted that in an eight-day trading period from July 29, 2011 to August 9, 2011 there were more than 395,000 transactions executed in SIGMA-X that violated the "trade-through" rule, which requires brokers execute trades at the best possible prices, regardless of what exchange it is on. The rule requires if there's a better price quote somewhere else, it has to be routed there for execution.
Goldman, FINRA says, was unaware that it was violating the trade-through rule (also known as the Order Protection Rule) during that period, but FINRA also notes the violations were not detected "in a timely manner."
Goldman returned $1.67 million to customers it had disadvantaged in those 395,000 trade-throughs.
FINRA also said Goldman failed to maintain written policies that were designed to prevent trade-throughs and didn't do regularly surveillance to determine if its policies were effective.
Goldman, as is typical in these types of cases, did not admit nor deny the charges, but consented to the entry of FINRA's findings.
What's all this mean? 395,000 trade-throughs is not a small number; assuming 100 shares a trade, that's 40 million trades, but let's assume an average of 200 shares a trade, that's 80 million shares. While the amount returned was not huge ($1.67 million), and this occurred several years ago, it's likely to get some attention because the orders were executed through Goldman's SIGMA-X dark pool.
Is this a system-wide problem? Are there similar issues with other dark pools? Will FINRA conduct a system-wide review looking for trade-through violations?
These dark pools (there are about 45 out there) have come under scrutiny recently, particularly after NY AG Eric Schneiderman asserted that Barclays misled users of its dark pool about who was trading in the pool.
No such allegations are made here, but any instances where dark pools are disadvantaging clients will attract attention.
One final point: Assuming those 395,000 trade-throughs were roughly 80 million shares, it would assume that Goldman was making about two cents a share on the deal. That seems to be about average these days. Surprised Goldman isn't doing better!