Interesting observation [1] on the lack of liquidity in this event:
"What's disturbing is the depth of the market during the event -- or rather, the utter lack thereof.
You can see the size map essentially disappear as the market's depth goes "cold."
One of the classical arguments raised by proponents of high-frequency trading and computer-based trading in general is that it "adds system liquidity." But if that was all it did there shouldn't be any impact on the market's price if that liquidity disappears.
That is, if HFT and computer-based trading merely add liquidity then there should be no directional bias in their decisions nor in the outcomes when it goes away.
But there are -- repeatedly. We don't see liquidity evaporate during "flash skyrocket shots" in the general sense. Instead, we see it happen during sharp downward moves.
The original purpose of the specialist system on the NYSE was to interject a human being in the trading environment who had a duty to buy when there were no other orders to do so, and to sell when there were only buy orders.
In exchange for the risk of getting caught glad-handed during severe events the specialist was given the privilege of making a spread the rest of the time.
Now we have computers that the operators claim provide the same function as the old specialists. But in fact they do not, because there is no duty associated with their actions.
We allow them to effectively front-run through physical location and latency advantages (that is, they see quotes and price before anyone else and thus can react faster than anyone else) and as a result during normal circumstances they effectively get to skim just as the specialists did. But they have no duty to remain online and provide that liquidity during severe market events as the specialist did and does."
>In exchange for the risk of getting caught glad-handed during severe events the specialist was given the privilege of making a spread the rest of the time.
No, they had old fashioned ways of withdrawing their liquidity. Not answering the phone, or pretending they can't see you across the pit.
>We allow them to effectively front-run through physical location and latency advantages (that is, they see quotes and price before anyone else and thus can react faster than anyone else)
That isn't front-running. Front-running means seeing the order flow before it reaches the matching engine and stepping in front.
>You can see the size map essentially disappear as the market's depth
When depth levels are taken out you have a great big empty space between the bid and the ask. It may have been 20 ticks wide yesterday (it was about 8 on eurostoxx, and those are chunky ticks). It's empty because someone has traded through those levels and taken all of the liquidity, so naturally it takes time for new orders to replenish the market. It isn't only because participants are withdrawing liquidity. Also, greater volatility means a greater dispersal of opinions about the correct price - so even when orders are put back in following the shock they are likely to be spread over a wider range of ticks. If you're a market maker (algorithmic or otherwise) who has just been smashed on the bid and seen the market drop 30 ticks, where do you put your new bid in?
The disinformation campaign by Wall Street against the HFT bogeyman is impressive. It isn't much different to record labels against MP3s or any other industry that is being disrupted by technology, but it is surprising to see a lot of the HN crowd falling for it.
We allow them to effectively front-run through physical location and latency advantages (that is, they see quotes and price before anyone else and thus can react faster than anyone else) and as a result during normal circumstances they effectively get to skim just as the specialists did. But they have no duty to remain online and provide that liquidity during severe market events as the specialist did and does.
The difference is that there are no regulatory barriers to entry. Anyone else who wants to is completely free to play this game - buy the computers and the connectivity and "see quotes and prices before anyone else". The comment suggests that HF market makers are taking risk-free profits, which is untrue.
There are news services that are dedicated to delivering important information that could affect markets as quickly as possible. I heard the news from one of these services directly before the market seriously dropped. The important question is why the markets reacted the way they did. There seems to be a sentiment that these sorts of moves are catastrophic and avoidable, but the reality is that these are movements reacting to real information from the environment that will affect the future state of economics. When bad news develops, more efficient markets drop faster in anticipation for this current or future event. This means that the information, or the bad event, has been diffused among the population.
With the AP situation, the news of such an extreme terrorist act would surely cause unbalance throughout Western economies; disrupting peace, supply chains, and many companies. This would have a direct impact on equities like the DOW. Within two minutes of the fake tweet, it was reported as being fake and the markets rebounded. In fact, I would guess that the majority of people hurt by this volatility spike were market making or algorithmic traders. They either were directly linked to twitter text feeds, selling on the way down and buying at the top of the retracement; or dump positions put on earlier in the morning. Of course, if the event happened, these traders would have been the winners.
The problem with this current situation is that the high-speed news delivery services spread this fake tweet without any verification (except trust that the AP feed is trust-worthy), and people with direct access to these services were most likely individuals with access to trading accounts, allowing them to drop a variety of assets, most notable the DOW or S&P.
> the news of such an extreme terrorist act would surely cause unbalance throughout Western economies; disrupting peace, supply chains, and many companies.
And that, in a nutshell is the whole problem. We should not react to such news at all, it would solve the terrorist problem and the economic issues in one go.
> We should not react to such news at all, it would solve the terrorist problem and the economic issues in one go.
That isn't how markets work and, more importantly, if you were to institute some policy that forced them to work that way, you'd get less liquidity. In my opinion, these shocks are a sign that the market is actually working properly.
As an aside, I know a lot of people who faded the move (remember, you have a choice!) and made quite a bit of money (smallest number I heard was $50k).
"In fact, I would guess that the majority of people hurt by this volatility spike were market making or algorithmic traders."
I think it's the other way around. I guess it depends on the volume on either leg of the spike, but the market makers who were buying on the way down would have been accumulating inventory at reduced prices and could have made a tidy profit when things bounced back. On the other hand, any trader with a stop loss order within the range of the spike probably got hurt by being forced into selling at a reduced price.
...but the market makers who were buying on the way down would have been accumulating inventory at reduced prices and could have made a tidy profit when things bounced back. On the other hand, any trader with a stop loss order...
So your claim is that market makers who were promoting an orderly market (by pushing the price back up) made money, while speculators who contributed to the problem (by pushing the price down) lost money?
Sounds like the market is at least getting the incentives right.
Yea, good point, it all depends on where the volume was. I would break it up this way:
The initial sell off: These orders would be executed from traders that use a live twitter feed to make buying or selling decisions. The underlying algorithm would have been able to execute these orders faster than any human. The serious news would have driven them to sell quickly, causing the market to start diving. Most of these probably made money, depending on how they were able to react to the news that it was fake. It is possible that there were algorithms selling all the way down the entire spike with the expectation that it was going to go down even further, losing money on most of the positions during the bounce. After-all, the bounce was faster than the drop.
Secondary sell off: While the algorithms are executing sell orders, market makers would begin to realize that a huge volatility spike was entering the market place, whether they heard the bad news or not. Because market makers train their models on relatively low volatile intra-day data, this spike would have been enough to reach either max positions or max downdraw, forcing the traders (or the market making algos they created) to cover positions for a loss on the way down.
Some interesting context/speculation from my submission on this yesterday [1].
"The Twitter data stream has been available to high frequency traders since at least 2009. That's when a company called StreamBase began to incorporate Twitter in the firm's "complex event processing" service, which is basically a platform that aggregates data from a vast array of sources for hedge funds and investment banks." [2]
This would seem to be a new way to monetize hacks. Go short on the market, trigger your fake tweet, wait a few seconds for the flash crash, and cover your short. You could make huge profits this way. I wonder if someone did in this case.
So what? As long as there is no evidence tying you to the hack. You could just claim it was a coincidence, that you were expecting the market to go down for other reason.
And then you get called in for lengthy interviews, and you get scared and make unrelated false statements, and get convicted of a felony. C.f., Martha Stewart.
Clearly it would be illegal. I was thinking about this as an extension of the market for day 0 exploits. Say you have found a new exploit. According to articles like this, you would typically sell it to some government spy agency, or even worse "bad guys", for something up to $500k.
This would be a way of converting that exploit into much more. If you had some experience in finance (or knew the right people), you could easily make millions on a series of trades. Illegal, sure, but this is for a person already used to dealing in very shady corners.
It would be hard for a "retail investor" like myself, but 30 seconds is a lifetime to the high frequency guys. I'm sure there is a pretty big group of people who would have no trouble doing it.
Acting quickly on news is a risk that they hope will pay a reward, but there is no guarantee that the news is correct and thus the risky position will be rewarded.
If you wait for verified, accurate news, then anyone who moved before got to move before you. Thus, participants in that market move when they want to accept the risk they want.
It doesn't matter if you verify Twitter with the best security around because markets will and always have reacted to rumor and unverified speculation and news. It's a gamble that the participants willingly make.
But there is much more money to be made in reacting to the news at millisecond speed, and therefore without any critical analysis whatsoever. Many of the orders that caused the drops were probably placed by machines.
Real-time semantic analysis-triggered trading seems to be what the kids are up to these days...
For a fun conspiracy theory (or at least, future conspiracy theory), consider that the "hackers" could have been paid by someone to take control of the AP's twitter feed and post something that might cause such a reaction. The temporary dip could make a lot of money for someone who was expecting it.
I'm not saying that's what happened here, but it wouldn't be surprising to hear about some fund or investment firm employing such a tactic in the future. Just imagine what would happen if organized crime got into the HFT game? That is, assuming they're not in it already.
You've only linked to print advertising and newspapers' declining revenue, i.e printing on dead trees. Printing on paper then selling that paper isn't a requirement for journalism to happen.
Apparently you didn't bother to click on them and view the image - the first one clearly shows online as a fraction of the total revenue for 2011 - sitting at around 10% of the industry's $24bn.
"What's disturbing is the depth of the market during the event -- or rather, the utter lack thereof.
You can see the size map essentially disappear as the market's depth goes "cold."
One of the classical arguments raised by proponents of high-frequency trading and computer-based trading in general is that it "adds system liquidity." But if that was all it did there shouldn't be any impact on the market's price if that liquidity disappears.
That is, if HFT and computer-based trading merely add liquidity then there should be no directional bias in their decisions nor in the outcomes when it goes away.
But there are -- repeatedly. We don't see liquidity evaporate during "flash skyrocket shots" in the general sense. Instead, we see it happen during sharp downward moves.
The original purpose of the specialist system on the NYSE was to interject a human being in the trading environment who had a duty to buy when there were no other orders to do so, and to sell when there were only buy orders.
In exchange for the risk of getting caught glad-handed during severe events the specialist was given the privilege of making a spread the rest of the time.
Now we have computers that the operators claim provide the same function as the old specialists. But in fact they do not, because there is no duty associated with their actions.
We allow them to effectively front-run through physical location and latency advantages (that is, they see quotes and price before anyone else and thus can react faster than anyone else) and as a result during normal circumstances they effectively get to skim just as the specialists did. But they have no duty to remain online and provide that liquidity during severe market events as the specialist did and does."
[1]: http://market-ticker.org/akcs-www?post=220071