LPS5 - Thank you for the thoughtful and intriguing response (and your comments on the other board regarding the same issues). Let me address one or two of your comments and throw in another thought or two to boot.
First, I understand VWAP has been around as a formal calculation for a long time now -- but from what I have been reading, the percentage of institutional block trades that are executed in a synthetic VWAP fashion, e.g., through time slicing intended to approximate a VWAP price (whether electronically, through POSIT, or other systems that offer time-sliced VWAP-like executions or manually by human traders), has increased by several multiples in the past few years. And the use of the derivative as a benchmark to measure traders is increasing along with it -- I am aware of studies by Elkins McSherry and Able Noser in the past year that attempted to measure trader performance against VWAP prices. Of course, there is an academic paper published in the "Education" section of the PHLX website that is a little dated, that addresses the issue from 3-4 years ago.
The Elkins McSherry study, which concluded that a VWAP price represented price improvement 83% of the time to those who trade "the regular way" -- even with time slicing techniques intended to disguise volume or approximate the benchmark -- suggests that those buysiders who think VWAP can be "beaten" are wrong a large percentage of the time -- which begs the question whether the psychological desire to be treated in a more hands on fashion and to do better than a derivative price that is by definition "average" through the exercise of cognitive decisionmaking is going to win out over price improvement in the long run? I don't know, but it is an interesting question.
As for the best execution (and the Potter Stewart analogy, which is priceless, BTW), you make a great point -- but again, in the context of an exceptionally large fill, where it is unrealistic to assume that the fill can be achieved in any way other than slicing it up in some fashion and timing it through the day, since there is rarely a single moment at which "best execution:" can be measured (as each portion of the total block is necessarily timed differently) doesn't it largely become a matter of looking at the averaged price achieved through execution of the block as an entirety, and comparing it to what was achievable for the block as a whole using other methods (putting aside -- only for the moment -- the effect that taking a large percentage of possible trades out of the call market and executing them at a derivative price does to the overall average)? I'm not saying that the Elkins McSherry finding is gospel, but it does suggest that comparatives can and should be measured, and that cognitive decisionmaking does not necessarily beat a homogenized price.
On your point regarding the skewing of the market when buy or sell interest is withheld from a continuous market, I have 3 observations. The first is that the SEC approval which granted the PHLX the right to operate eVWAP as a facility of the exchange actually placed a 20% limit on the percentage of total shares in any one security that could be traded on eVWAP before lights would go off at the SEC and heightened scrutiny would be triggered to see if the concentration of volume being traded over eVWAP in that security would be more prone either to manipulation or skewing (I don't think the word "skewing" was actually used, but I am quite sure the 20% threshold is there -- not as an absolute limit, but as a point at which the SEC would want to look more closely to ascertain the effect of that much concentration). Second, given the nature of the eVWAP system, it is impractical to think that true large block traders (say, for the big managers who are trying to move 1000 trades with an average volume of 100,000 shares each) will be able to unload all but a relatively small percentage of that total volume in a system like eVWAP -- I think its real appeal is as an adjunct way to diminish in part the effects of market impact which, for better or worse, are exacted on the large block trader, particularly with decimilization making it easier to front run a large order and, as you say, performance being measured in basis points. If I can unload 20% of my fill using eVWAP, then I still need to put 80% of my order into the call market, which achieves (albeit to a less dramatic extent) the effects of the market's reaction to evident buy or sell interest -- but 20% of my position has been filed at a price which may represent price improvement 83% of the time. Also, since eVWAP is a matching system, it is pulling buy and sell interest out of the call market in equal proportions, which presumably reduces the skewing effect you describe. still, I agree that if a very large percentage of institutional fills were made through any off-market derivative pricing system like eVWAP, the absence of that trading in the call market would necessarily have an effect on the trading that remains in that market. Third, and this relates to the 2nd point, the existing call market system in effect on the NYSE and NASDAQ which allows specialist (and in the case of the NAZ, the market maker) the ability to see all trading activity as it enters the market in exchange for the specialist's or MM's agreement to assure an orderly marketplace (and provide fills where there are order imbalances) gives the specialist a huge advantage in terms of playing the trading activity that it sees into subtle price improvement that benefits the specialist. I'm not suggesting it is improper, as the specialist may have to sell when selling makes no sense to meet its obligation to ensure an orderly market, which is a risk and exacts a cost -- but as the specialist is in the business of making money, it is natural to assume that it will take advantage of its superior knowledge to improve its income from trading, which intentionally or unintentionally also has some degree of "skewing" effect. Clearly, techniques used by MMs (such as purchasing order flow to facilitate internalization), or taking take stock into inventory (or ridding oneself of inventory) because of the perception or order imbalance that only they can see, are utilized in many instances to the disadvantage of the customer. Presumably, in an efficient market, the Specialist or MM which is too "greedy" in its approach to this activity will lose out as customers seek better executions elsewhere (and with stocks that trade 50 or 100 MM shares a day, the action moves too fast for this to exact a cost that is material to the overall transactions) -- but with many players whose trades are too big for the market needing to spread it around, and a system whose basic structure gives one group a better vantage point from which to see the market as a whole, it still seems as if the potential for profit at the expense of the customer needing the fill (and thus the skewing of the market in the direction adverse to the large order) is acute and is pushing the large block traders to seek relief in more hospitable venues. In my mind, the chief advantage of a system like the eVWAP system is not so much the pricing itself (though the use of an intrinsically fair price is pretty important), as the fact that the trade takes place outside of the call market, where the counterparty really doesn't gain knowledge through the trade itself from which it can benefit at the direct expense of the party with whom it is matching. But I quite agree that the system would not work well if too much of the market ended up there.
One last point -- you suggest at the end of your response that "Sticking near or at the pricing benchmark - whether VWAP, the new "RPM" system, or whatever - is "good enough," putting the guideline - rather than the buysider's investment objectives, the trader's market savvy, or other desk attributes - in the driver's seat, so to speak" winds up with us having "the tail . . . wagging the oft-cited dog." While I understand the notion that market savvy or desk attributes may make for some traders who simply beat the market more often than not (that 17%) -- relative to passive matching systems which, even though they achieve a better price as much as 83% of the time, do not think or act strategically -- I do not think it is counter to the buysider's "investment objectives" to use systems that benchmark prices -- the choice to trade in or out of a stock today (as opposed to tomorrow or last week), is almost as important as whether I time slice or trade 1/2 hour after the open, or based on what I am seeing -- and, moreover, with some institutional traders, particularly index funds and shadow index funds (dealing with fund placements or redemptions), choice does not come into play, and they are forced to trade at inopportune moments in order to achieve the fill (which is, by necessity, the only objective they must adhere to). For those users, any vehicle which minimizes market presence while they go about achieving their objectives in the face of market forces moving the other way would seem pretty important.
Anyway - it's all quite interesting -- thanks again for taking the time to respond.
MST |
