How feasible is a kill switch?
Even before the SEC's roundtable on technology and the markets, the idea of a kill switch was gaining ground. In the wake of the roundtable, one could argue that the idea has generated near-consensus backing.
So what exactly would a kill switch do? The idea is to set up some software that would be able to shut down a member's trading if the trading is somehow out of whack. One NYSE executive was quoted by MarketWatch saying, "We believe that a kill switch-like solution is may be the most widely agreed-upon solution that may be implemented in a reasonably short time frame."
Actual implementation of kill switches by exchanges is not likely to happen immediately, and there are some thorny questions about how the kill would actually work. There's a chance that the kill software itself may not work as advertised. It would be the ultimate irony to have software bugs that render kill switches, which were designed to combat algo bugs, inoperable.
What the exchanges need to do is publish in excruciating details when the kill switch will be invoked, laying out clearly all the triggers. At the roundtable, many expressed the idea that simplicity is better. Other said that firms ought to be given a chance to fix the problem before the kill switch is used. Still others would want companies, not exchanges, to control kill switches.
The biggest danger is that the kill algorithm ends up hampering legitimate high-frequency trading. There would appear to be a lot of pitfalls, but it does appear that the idea is moving ahead.
Elusive solutions to technology glitches