While the defenders of HFT continue spouting their usual platitudes (with the latest piece of "anti-hyperbolic" fluff coming from "Mr. Quant" (but don't call him an HFTer) Cliff Asness himself who said overnight that "markets are "rigged" in favor of, not against, retail investors"... so - rigged?) the reality is that while one can debate the ethics of HFT frontrunning orderflow until one is blue in the face (or until Goldman tells the DOJ to slam the hammer on the high freaks once and for all), the biggest adverse impact from HFT continues to be the inherent instability that algo trading creates in the market.
One needs only to recall what happened on May 6, 2010 and extrapolate that nothing since then has changed, and in fact HFT has become even more ubiquotous and pervasive, to get a sence of where the rigged, and exceptionally frail and broken market is currently.
For empirical evidence of just this, we once again go to the usual source which everyone ignores until months after the fact is seen as having been right about everything, Nanex, which looks at one particular aspect of market instability, namely Limit Up, Limit Down circuit breakers and finds something very disturbing:
The number of stock halts from the new Limit Up, Limit Down circuit breaker (LULD) is on the rise. Some of this increase is due to multiple halts in the same stock on the same day, and a significant subset of these multi-halt events are triggered, not from actual trading, but from a mere widening of the quote spread. One stock exemplifying this is Collabrium Japan Acquisition Co. (symbol: JACQU, market cap: $55M). On May 20, 2014, the stock halted 19 times! Even more interesting, every halt came exactly 5 minutes and 15 seconds after the previous halt. More interesting still, it halted at the same exact times the next day!
Keep an eye on the increasing incidence of these critical events which simply show just how broken the "market" is because as Nanex concludes, "LULD was created in response to the Flash Crash, and the creators of this method better hope there is not another flash crash, because LULD will be a disaster in that environment."
A table showing JACQU's halts:
# May 20, 2014 May 21, 2014 Halt Time Time Between Halts (mm:ss) Halt Time Time Between Halts (mm:ss) 1 09:45:15 n/a 09:45:15 n/a 2 09:50:30 05:15 09:50:30 05:15 3 09:55:45 05:15 09:55:45 05:15 4 10:01:00 05:15 10:01:00 05:15 5 10:06:15 05:15 10:06:15 05:15 6 10:11:30 05:15 10:11:30 05:15 7 10:16:45 05:15 10:16:45 05:15 8 10:22:00 05:15 10:22:00 05:15 9 10:27:15 05:15 10:27:15 05:15 10 10:32:30 05:15 10:32:30 05:15 11 10:37:45 05:15 10:37:45 05:15 12 10:43:00 05:15 10:43:00 05:15 13 10:48:15 05:15 10:48:15 05:15 14 10:53:30 05:15 10:53:30 05:15 15 10:58:45 05:15 10:58:45 05:15 16 11:04:00 05:15 11:04:00 05:15 17 11:09:15 05:15 11:09:15 05:15 18 11:14:30 05:15 11:14:30 05:15 19 11:19:45 05:15 Missed!
1. Daily Count of LULD Stock Halts.
Is FINRA or the SEC paying attention?
2. LULD Stock Halts by Date and Time of Day.
The size of each bubble indicates the halt number in a stock on the same day (the first halt would be tiny while the 10th halt would be a large bubble). A stack of progressively increasing bubbles shows us where multiple halts occurred in the same stock, such as the JACQU example in the table above: JACQU's halts on May 20 and 21 appear as two stack of bubbles at the lower right corner of the chart below.
3. Zoom of Chart 2