Hi Eric,
Would it be possible to expand on these examples, for instance, how could the flash order lead to a price improvement and how would you become aware that you were being 'front run'?
The flashed order can lead to price improvement in two ways: You could get filled at a better price, between the bid and ask. More likely, though, you would get the same fill that you'd otherwise get, although at a reduced ECN fee than you would pay if you hit the posted offers directly (for example, hitting an BATS offer using the BATS route). Without researching the specific fees to get a 100% accurate example, I'll pull a few numbers out of thin air, instead.
Assume you want to buy 500 shares of a stock with a bid/offer of 24.14/24.15. You see 2300 shares on the offer at 24.15, with 1000 shown on BATS, 500 on ARCA, and 800 on Nasdaq. If you place a Nasdaq, BATS or ARCA order to buy 500 at 24.15, you will likely get you full order filled at your order price, but if using a direct access broker, you will pay an ECN fee of perhaps 0.3 cents per share, or $1.50. By sending the order through EDGX and flashing the order, you will likely still get filled with the same 500 shares at 24.15, but instead you might only pay an ECN fee of 0.1 cps (ballpark, depending on where your order got filled), or $0.50, for a saving of a buck. Many traders might not care to save a buck, but if you are making hundreds of orders per day, it can really add up.
In terms of front running, I believe (but can't prove) that those seeing these flashed orders will sometimes act against your best interest when they see an advantage for them to do this. For example, assume that the stock is priced at 35.18 bid and 35.24 offered. The offer only has 400 shares, and the next offer is 100 shares at 35.27, then 200 shares at 35.32. Let's assume that you place a flashed order to buy 1500 shares at 35.24. => I suspect that a disproportionate amount of the time, you will not get a fill for even the first 400 shares that are displayed on the books. Instead, one of those flashed destinations that got an early peek at your order will instead choose to immediately execute to buy those 400 shares for their own account, hoping that your need to buy 1500 shares will enable them to sell those shares back to you (or someone else) at a few cents higher price. Again, although I can't prove this, I do know that my slippage averages for flashed orders compare favorably to non-flashed orders for thick, liquid stocks. Alternatively, the flashed orders tend to suck, on average, when used to buy more thin, illiquid stocks. Frontrunning (or whatever term you want to use for this) is a logical conclusion as to the cause, IMO. |
