February 09, 2004

Stock Trading Attacks

Zooko O'Whielacronx wrote:

"Your paper about mutual funds, which I've only begun to look at, reminds me of an attack on a "stock market implementation in E." In my own opinion, my crack was trumped by Ralph Hartley's generalization ."

Ralph Hartley's generalization (read here) is spot on, with the caveat that the attacks apply equally to any stock exchange, rather than Mark Miller's demonstration stock exchange in particular. That's because these attacks are based on the coordination problem, and the timing aspects of stock trading dynamics. In that sense, they perfectly mirror some of the attacks that are going on in the mutual fund world today.

In the testimony/paper of a few weeks back, we said that RTGS solves these problems, although that's a generalism, and depends on the assumptions. The really important point is that the shorter the timeframe for the trade settlement, the shorter the opportunity for abuse; and the financial cryptographers in the above links came to the same conclusion.

Which makes it all the more poignant that the trading world managed to get to T+3 settlement, and stopped there. Desire to move to T+1 (we are talking one day here, not minutes or seconds) is non-existent, and opportunity for fraud remains as present as it ever was - given that most of the frauds of the mutual funds industry occurred within a one day timeframe.

Posted by iang at February 9, 2004 10:09 AM | TrackBack