High Frequency Trading: The Best Of Technological And Financial Innovation... Or The Next Bubble?
from the where-are-things-headed dept
Over the last couple years there's been growing chatter about the rise of "high frequency trading," which is the increasingly sophisticated algorithmically-driven way of financial trading, where it has little to do with how smart your investment philosophy is, but how fast your hardware and algorithms run. As a principle, there's nothing wrong with the concept of high frequency trading. And, as many defenders of the concept point out, such systems, in theory, provide more liquidity to many markets, and basically skim pennies off the top in return for that liquidity. The potential problem, however comes in when such operations take over the market. The latest estimates put high frequency trading at 61% of the market -- up from 30% just five years ago.That should be a warning sign. It's typical, but you can see it in plenty of previous Wall Street meltdowns as well. After someone figures out a "system" for making lots of money (say, mortgage-backed securities a few years back), everyone starts piling in. Then, the "innovation" occurs. Now, much of it is well-meaning, and even useful. With mortgage-backed securities, things like credit default swaps actually were a very useful insurance tool originally. But at some point, they basically flipped from insurance to gambling. People weren't using them to back up an investment, but as the investment itself -- so you'd actually have what was, in effect, thousands of people all buying an insurance policy that one house wouldn't burn down. If that house burned down... the insurance company (hi, AIG) defaulted, and everything comes crashing down. The problem is that these systems become so complicated that it's actually pretty difficult to figure out what the "trigger" is and how the disaster will spread. No one accurately predicted how the last Wall Street meltdown would occur (though some certainly predicted a meltdown), and the fear with the rise of high frequency trading is that the situation is even more opaque. What's happening is built into the algorithms, and with more and more companies piling in, it's inevitable that some of those algorithms are going to have a bug (or, not even a bug, but basically programming to do something that has serious unintended consequences).
Again, I doubt there's anything nefarious in most of this (unless you consider making money nefarious -- which I don't). But, at some point things get overwhelming, and many are beginning to wonder when we reach that point. I'm all for financial innovation and technology innovation -- but I have to admit to a bit of worry when the tech innovation seems to be taking over to such a level that there's little rationale for the financial side. It's about who has the better techies and hardware, rather than who has the better financial thesis, and that leads to dangerous results, because the purpose of the market is separated from the mechanisms that make the market run. When you get that kind of separation between form and purpose, bad things happen.
Thank you for reading this Techdirt post. With so many things competing for everyone’s attention these days, we really appreciate you giving us your time. We work hard every day to put quality content out there for our community.
Techdirt is one of the few remaining truly independent media outlets. We do not have a giant corporation behind us, and we rely heavily on our community to support us, in an age when advertisers are increasingly uninterested in sponsoring small, independent sites — especially a site like ours that is unwilling to pull punches in its reporting and analysis.
While other websites have resorted to paywalls, registration requirements, and increasingly annoying/intrusive advertising, we have always kept Techdirt open and available to anyone. But in order to continue doing so, we need your support. We offer a variety of ways for our readers to support us, from direct donations to special subscriptions and cool merchandise — and every little bit helps. Thank you.
–The Techdirt Team
Filed Under: bubbles, efficiency, high frequency trading, liquidity
Reader Comments
Subscribe: RSS
View by: Time | Thread
Having It Both Ways?
versus
If there’s “nothing wrong” with something, then what’s wrong with it becoming more successful, to the point when it “takes over the market”?
[ link to this | view in chronology ]
Re: Having It Both Ways?
It can be useful on the margins to pump in extra liquidity. But when that's all it's doing, then it's not about adding value (liquidity) to the market, but just about trying to move money around.
It's the difference between transactions that add value and transactions that simply try to capture a slice of value.
[ link to this | view in chronology ]
Re: Having It Both Ways?
But so what? Isn’t the measure of market “success” the fact that you can make a profit from an activity? And therefore, that the more of that activity you can perform, and the more profit you can make, the more “successful” you must be?
Oh, they do indeed add “value” to their transactions all right—they get their cut every time the money goes round. You don’t think they’re doing this out of the goodness of their hearts, do you?
[ link to this | view in chronology ]
Re: Re: Having It Both Ways?
Arbitrage is not adding value. It's just redistributing it.
[ link to this | view in chronology ]
Re: Re: Re: Having It Both Ways?
Arbitrage is not adding value. It's just redistributing money.
[ link to this | view in chronology ]
Re: Having It Both Ways?
What exactly is your problem with that? Efficiency in markets is all about efficiency in using available information, is it not? That’s all arbitrage is doing. So why should it be wrong, when efficient markets are not wrong?
[ link to this | view in chronology ]
Re: Re: Having It Both Ways?
[ link to this | view in chronology ]
Re: Re: Having It Both Ways?
[ link to this | view in chronology ]
Re: Re: Re: Having It Both Ways?
I have no problem with fast computers or even fast trading: faster the better for markets -- but looking into someone else's limit price might be unethical / against the law!!
[ link to this | view in chronology ]
Re: Having It Both Ways?
Lawrence, please. Too much of *anything* is a bad thing, and you know that.
[ link to this | view in chronology ]
hedging vs. speculation
Mike: I'm sure you of all people know there's no fundamental difference between hedging and speculation. At what point could things have flipped?
That said, I like the form/purpose sanity check -- it's simple and broadly applicable.
[ link to this | view in chronology ]
Re: hedging vs. speculation
You could say there was more risk than needed prior to the trade, and call that speculation. But not in general. For example I don't count growing a crop as speculation, and hedging trades can be useful there.
You could also claim one side of the trade must have higher risk if the other side has lower. But this is also not generally true. It is better to think of hedge risks as anti-correlated, not higher or lower. Two complementary portfolios can have lower net sensitivity to the risk being hedged.
[ link to this | view in chronology ]
Re: Re: hedging vs. speculation
Growing crops is a productive activity. Playing the stock market isn't.
[ link to this | view in chronology ]
Nefarious? Well, sorta...
But shouldn't making money come from economically useful activities? The money Wall Street is making via high frequency trading (HFT) is not because they're generating new value or durable economic outcomes -- they're just figuring out how to game the system for their own advantage.
The only "value" created in HFT is keeping competitors out of the marketplace long enough to get a microsecond leg up on a trade. But it gets worse. With that first-mover advantage, they also create the illusion of market momentum and value perception when there's no real basis for it. Stock prices then move because they moved. That's not really new, but now it operates well beyond the level of human perception.
So making money in this context IS nefarious, at least to me. Because I can't compete in that market and neither can you, no matter how hard we work or how smart we are. The number of companies that get the privilege of "competing" is limited to the number companies that already have control of the system. Furthermore, their actions in this non-market have repercussions that reverberate through virtually every market in the rest of the economy.
To me, that's fundamentally wrong, in a moral sense. We effectively have a royalty court of self-dealing financiers within a society that defines itself as the "land of equal opportunity." And since it's being done consciously -- with clear understanding of the consequences in advance (if I can see it, you know Goldman Sachs knows full well what's going on) -- then I gotta go with the "nefarious" label.
[ link to this | view in chronology ]
I disagree.
I don't claim to completely understand how high-frequency trading works, but my impression was that basically these companies' computer systems would note that someone (i.e. a slower computer) is trying to buy a bunch of shares of some stock, and then in a tiny fraction of a second buy up all available shares of that stock, allowing them to (another fraction of a second later) sell shares for slightly more to the actual buyer.
I really don't see how doing that is adding any value and it seems like if anything it would decrease liquidity. It basically just seems like a high-tech analogue of insider trading.
The government should simply put a tiny tax on all financial transactions. It would have basically no effect on ordinary investors, mutual funds, etc., would act as a small tax on day-traders, hedge funds, etc., but would bury worthless scams like high frequency trading.
[ link to this | view in chronology ]
Re:
I don't think that's quite it. I think by the time anyone, even these centrally located computers, can tell that someone else is "trying" to buy a stock, it's too late (correct me if I'm wrong). I think what these programs do is look for other events that indicate people (somebody, they don't know who) will want to buy a stock. Then they buy it before those others have a chance to.
I think the liquidity theory is that it makes more shares available from a single seller than would otherwise be the case. I don't know if that theory holds water (ha, see what I did there?). I'm not sure we can really label it cheating. Is it wrong to make trades faster than your competitor? How slow are we going to require them to go? Rather than banning it, I think there should be some kind of incentive to limit its influence on the market. 60% sounds way too high to me, and what if it keeps growing to 80 or 90%? It sounds like (I know, uninformed rambling) a(nother) disaster in the making.
The government should simply put a tiny tax on all financial transactions. It would have basically no effect on ordinary investors, mutual funds, etc., would act as a small tax on day-traders, hedge funds, etc., but would bury worthless scams like high frequency trading.
That just might be the way to go. Or maybe have a tax apply to stocks that are held for less than a certain amount of time. Is there some threshold below which any trading is likely to be arbitrage rather than an activity useful for the market? Say 5 minutes? Even day traders wouldn't be affected by that I think.
[ link to this | view in chronology ]
Re: Re:
Consider the following example: I want to buy a house and put in a bid. If the seller peeks at my bank book and now knows how much I can afford - he keeps rejecting the bid till I offer everything I have in my bank. This peeking prevents a true negotiation from occurring and is unethical.
What is worse is that a third party peeks at my bank book, buys all houses in the neighborhood (for a few nanoseconds or whatever) and makes me pay my last price! This third party, besides being disruptive, unnecessarily inflates the price of the asset!!!
In a reverse scenario, the disruptive party over-deflates the price of the asset.
Liquidity? Rubbish!!! Some of these algorithms may just be adding Volatility -- that only benefits short time traders not the markets and definitely not the society. No net value added to the society.
If we have fast computers, we better come up with more ethical algorithms - now that is competition worthy of praise.
[ link to this | view in chronology ]
Re: Re:
Some quick Googling finds this explanation of how high-frequency trading works; the further details make it seem if anything worse than what I described above.
Just to be clear, while I think high frequency trading is horrible for the market and basically just makes things more expensive for legitimate traders, it seems unlikely to generate another "bubble" problem (except if financial firms become so reliant on the profits from high frequency trading that they leverage themselves too much and then suddenly the government and/or the stock exchanges finally get around to fixing the problem).
[ link to this | view in chronology ]
Re: Re: Re:
"While markets are supposed to ensure transparency by showing orders to everyone simultaneously, a loophole in regulations allows marketplaces like Nasdaq to show traders some orders ahead of everyone else in exchange for a fee."
"The disadvantage was that the "algos" had engaged in something other than what their claimed purpose is in the marketplace - that is, instead of providing liquidity, they intentionally probed the market with tiny orders that were immediately canceled in a scheme to gain an illegal view into the other side's willingness to pay."
"HFT systems that front-run are able to garner risk-free profits. This is in fact the reason such a practice is banned - their "risk-free" profit is your guaranteed loss. Remember, the markets are in fact a negative-sum game (due to trading costs) - if there is a "risk-free" opportunity out there it can only exist because someone else is guaranteed a screwing."
It really does seem pretty bad, even if it doesn't lead to another collapse.
[ link to this | view in chronology ]
Whats wrong?
[ link to this | view in chronology ]
Re: Whats wrong?
Speed of trading and related inventions may just be good for overall market as things become more efficient. Problem is that there may be an unethical practice in some of these algorithms.
Consider the following example: I want to buy a house and put in a bid. If the seller peeks at my bank book and now knows how much I can afford - he keeps rejecting the bid till I offer everything I have in my bank. This peeking prevents a true negotiation from occurring and is unethical.
What is worse is that a third party peeks at my bank book, buys all houses in the neighborhood (for a few nanoseconds or whatever) and makes me pay my last price! This third party, besides being disruptive, unnecessarily inflates the price of the asset!!!
In a reverse scenario, the disruptive party over-deflates the price of the asset.
Liquidity? Rubbish!!! Some of these algorithms may just be adding Volatility -- that only benefits short time traders not the markets and definitely not the society. No net value added to the society.
If we have fast computers, we better come up with more ethical algorithms - now that is competition worthy of praise.
[ link to this | view in chronology ]
[ link to this | view in chronology ]
Analogy time.
Almost everything in life is about balance and timing, economics are no different.
Now if you take the subprime example as a tale of caution, things got out of control because wall street was very good at hiding the toxic assets diluting them into others good assets, so nobody actually saw it coming.
http://www.physorg.com/news180640677.html
High Frequency Trading means that economists and accountants will be learning FFT and wave transforms? LoL
[ link to this | view in chronology ]
Re: Analogy time.
Like rape is good, in moderation.
[ link to this | view in chronology ]
event
[ link to this | view in chronology ]
Re: event
..so, shove it up your ethical compass.
[ link to this | view in chronology ]
The idea is to make an algorithm that *learns*, so you have it learn to predict what the stock market will be like next. Of course, that is pretty much impossible with human players because they have all these external factors that affect their decisions (news, emotions etc) that make things quite random. But, if all other players get replaced by algorithms, things suddenly become a lot more "deterministic" (predictable). It becomes a race of who has better algorithms that can learn the behaviour of other algorithms, that in turn learn the behaviour of other algorithms (including themselves) and so on.
Humans perform better against other humans, and such algorithms only do "ok" in a real situation some of the time. Now, if humans are phased out completely, things will be very different... and that could be a problem. A bubble forms when investors keep doing the same thing because it gets them short-term gains, and ignore common sense -- these algorithms can't have long-term plans and have no common sense, so they are built to do whatever action will get the greatest short-term gain. If they are left unchecked for too long, my feeling is they are very likely to implode, like many machine learning algorithms do when pitched against each other in similar situations (best case, they just start doing stupid, random things). But that's just conjecture.
[ link to this | view in chronology ]
Re:
what you describe could be the normal hurdles in the evolution process. There are bound to be mistakes (and therefore some species, i.e. paths chosen, become extinct) and some that work. If machines lack common sense, then they must be taught common sense and that will happen through evolution: initially some bad machines will have to come around that pave the way for better ones. (On the same note, we come around some humans that lack common sense too - our brains are only slightly different from the computers. BTW: common sense is very dependent on context).
Again, in the same way - a machine's ethics and values are only as good as its programmer's - so that will have to be built into them: and the citizens (you and me) will have to demand for such propriety.
Good thing can be that we don't need to teach fear and sheep mentality to the machines - as those seem to plague the human driven markets.
[ link to this | view in chronology ]
Re: Re: Re: (nasch / Dec 24th, 2009 @ 1:01 pm)
I must have read these articles on the web when I posted the two earlier posts.
[ link to this | view in chronology ]
response HFT
On the technology front - we at Intel are now being asked frequently for expsoure to the latest CPUs - where these can give a nanosecond advantage in hitting a price. This demand is coming from venues as well as traders - it is a market. Each year we are bringing out perfomance features the percentage improvement depends on how good the code is - but is rarely less than 10%. The point is, HFT is the Formula 1 of trading - and if you are in it it is about focus and technology engineering to ensure that you are at the front of the game in all respects.
[ link to this | view in chronology ]