Just one slight clarification, Mark:
VWAP orders on the first iteration of VTS will be matched before the opening bell, anonymously, using a matching algorythm based on order placed during the previous 16 hours or so through a computer network (available through the internet, intranet, B/D dedicated machines, 3com telecommunication devices, etc.) and secured through state of the art encryption (ATED devices) similar to that used to secure nuclear missile silo's. There is absolutely no human intervention - if two 400,000 share orders are placed on opposite sides of a security, those two big blocks may be matched against each other (e.g., the matching is not necessarily in small chunks, and all of it takes place before the opening bell). If no dedicated buyers and sellers match, B/D's have the capacity to step up and match the offsetting side of an order - they make those commitments up front (e.g., they have open ended commitments to buy xx,xxx shares of Merck if no other match with priority in the algorythm appears) which is intended to add liquidity to the system - critical to its success. The belief is that ATG has had considerable interest shown by B/D's in providing liquidity on this basis, in order to obtain liquidity and become enabled to trade NYSE listed stocks, previously the domain of NYSE floor specialists. If this belief is true (ATG will make NO official pronouncements, owing to its guaranty of anonymity), ATG will dwarf Optimark's system, which is based on parties filling in trading preference questionnaires in which trading parties reveal their trading preferences (e.g., I will buy xx,xxx shares more if the price rises by 5% in the first 2 hours, yadda yadda yadda) anonymously, though a securitized and fully automated system - the chief difference being that VWAP is an easy to understand, industry accepted pricing standard, whereas nobody fully understands Optimark's pricing methodology, which in my opinion is why its liquidity has been so woeful.
Traditional VWAP executions have involved institutions asking traders to slice their orders into 40 10,000 share slices, and trading it through the day - as Mark points out, while the method will approximate VWAP, it also lends itself to frontrunning due to the elements of human intervention and greed, so that B/D's make more $$$ at the expense of the institutions placing the orders - in the opinion of academics, between 8-12 basis points, which, over the course of a year of trading for any large institution, amounts to hundreds of millions of dollars in increased execution costs.
Hope that helps.
MST |
