<<<A simple example would be to buy 5,000 shares at the market every 30 minutes until the
order was filled.>>>
This is the point I was trying to get at. If buy orders are sent at market to obtain VWAP, then eVWAP should be superior, since its average is obtained from a different set of prices. Not saying eVWAP is sliced bread, just that relative to time slicing, eVWAP should result in a better price, since its average includes both executed sales and buys in a security.
In your example, assume you had orders of $65,000 shares from two clients to time slice and the spread was always 1/8 between the bid and ask. One wanted to buy and the other wanted to sell. Given the assumption the spread is always 1/8, the bid and ask price really don't matter, except that having numbers might make it easier to compare eVWAP to time slicing to approximate VWAP. For this discussion, how about 100 and 100.125 for the bid and ask. For each time slice order, it would seem that $8,125 would be paid to the market maker. The total funds received by selling at the vwap would be $6,500,000, and the total funds necessary to acquire at the vwap would be $6,508,125.
If that is accurate, then it provides a basis to compare eVWAP. Assuming there are no other trades, then the eVWAP price would be $6,504062.5. If this is accurate, then eVWAP represents a price improvement of $4,062.5 compared to the price that would have been obtained trying to use time slicing to obtain VWAP.
Agree? Anyway, the point of this is to see if there is a point to be made here or not. If this is accurate, and time slicing is more than a concept, then I score this as a plus for eVWAP. Over a large volume of shares, the saving could be significant relative to time slicing. |
