Ripple! Goldman Algo vs. Knight Algo


“You who choose to lead must follow
But if you fall you fall alone,
If you should stand then whose to guide you?
If I knew the way I would take you home.”

            Ripple – Grateful Dead

Just over a year ago, Knight Capital – a broker dealer and market maker – was brought to its knees by a runaway algo in the equities market that cost the firm hundreds of millions, as well as its independence. The trades made by that algo were forced to stand.

On August 20th – just over one year later, Goldman Sachs – a broker dealer and a market maker – also had a runaway algo, but theirs was unleashed on the options market. You can also read more about the options market SNAFU in this Bloomberg Article. Over one million trades were placed in a 20-minute period, resulting in trades taking place substantially outside of the prevailing market. The losses from these trades is not disclosed – but suffice it to say it is very, very large – and perhaps larger than the loss resulting from Knight’s trades of August 2012. Goldman’s trades will be busted. From the WSJ:

After hours spent reviewing thousands of transactions, U.S. options-exchange officials on Wednesday decided to cancel most of the trades caused by a technical glitch in a Goldman trading system, people close to the matter said.

That could reduce potential losses for Goldman but hit a broad swath of trading firms that were on the other side of those transactions.

We wish to highlight three issues this morning.

First. Isn’t there a collaborative policy by the exchanges called Clearly Erroneous?  Apparently there is for equities, but not for the huge options and derivatives markets. Also from that WSJ article:

Officials at exchanges such as NYSE, NASDAQ, CBOE, and ISE spent Wednesday sifting through each errant transaction and informing traders which ones would be “broken,” or canceled. Each exchange has its own procedures.

Second.  Isn’t there an SEC regulation called Reg SCI, which is supposed to regulate technology in the markets, and insure that those technologies are robust, so as to protect investors from “glitches”? Yes, there is. But first of all, Reg SCI is still being debated, and much of the industry feels that imposing extra burdens on exchanges will not help the SEC safeguard the markets with reasonable cost. Second of all, Reg SCI applies to exchanges, but does not apply to high frequency trading done by independent players and broker dealers.

We wrote a note to you just over one month ago that highlighted a comment letter from R.T. Leuchtkafer on Reg SCI. His comment letter made this exact point:

“It would be peculiar if Reg SCI at adoption didn’t apply to at least some segments of the high frequency trading industry: more peculiar still because by at least one measure Knight’s problems had a worse effect on confidence and investment than the May 6,2010 Flash Crash.”

Finally, why is there a disparate treatment with regard to the busting of trades? Goldman Sachs’s errant options market algo get bailed out, but Knight’s errant equity market algo did not. Perhaps we don’t have all the facts. Perhaps we do not know just how large this error was.

But perhaps the disparate treatment results from the fact that the derivatives market is still hugely profitable for the exchanges. Over half of their revenue today comes from these derivatives. In 2012, at the time of the Knight algo, a much smaller amount of exchange revenue was derived from the cash equities business. Some even consider cash stock trading a throw-away business for stock exchanges – which was their original purpose.

As we see the exchanges decide to bust the Goldman trades in their most profitable business side, we wonder if stock exchanges would have busted the errant Knightmare trades if those trades took place in 2008 or 2009, the heyday of equity market HFT. Perhaps this is an issue of business interests and not one of fair and consistent principled-based rule making.