My understanding is that HFTs function in a complete race to the bottom, with the only profitable activity being to trade faster and with slightly faster information than the next HFT. There is no secret sauce to keep them from eating each other as the costs to create a new HFT are not all that high vs. the profit opportunity of shaving a few percent off the commissions of the dominant trading system. Faster links and faster trades between exchanges mean they carry less risk when carrying out arbitrage (while also lowering the amount of arbitrage available).
The fact they made crazy profits initially is more of an artifact that they were competing against humans rather than other automated systems. In a few years we may see them reach the point of diminishing returns where all exchanges share a common pool of liquidity that adapts in a few millis to new information.
One explanation I've heard for HFT's profitability (which I'm too lazy to source at the moment) is that it's also an artifact of the granularity of prices permitted by the exchange, just like the 1/8 limit.
Instead of competing to save you a few more slivers of a cent per share, they compete to fill your order a few microseconds faster (at the next-highest price permitted). Seems most traders would prefer the money.
It really depends on how you define HFT. If we define it as the author does, then it's not so much a dead end, rather it is really hard to move in on. I like to think of latency arbitrage (including pairs trades, exchange arb, etc.) as a utility. It takes a massive amount of investment, but if you do that you can get a very steady return. The problem is that in most of those trades there is no money in being the second best. These leads to an environment where the people already invested are in an arms race to keep at the front.
If on the other hand, you are talking about other forms of algorithmic trading, there is still money to be made by being smart, having insight, having better procedures, or faster time to market etc. The arms race in latency arbitrage has actually made a lot of this easier as it has dramatically brought down the cost of entry to the non-bleeding edge of latency use cases.
I thought part of the problem is that exchanges charge more for faster access (e.g. physical collocation), so the exchanges profitability depends in part on creating a bidding war between HFT companies?
None of this is simply "arbitrage" as HFTers want the public to believe. The major brokers are perfectly capable of buying and selling at the best prices across any non-darknet market with even rudimentary order routers. What's happening is HFTers are gaining informational advantages and then exploiting them in microseconds. Based on realtime orders, they'll infer that someone wants to buy share abc at price x, deviation y, and then bet on that happening. HFTers can even infer which brokerage firms are placing the orders based on technical data including lag times and data-sharing partnerships.
HFT capitalizes on a kind of time arbitrage. They take the data from one market, and then transport that data using specialized, custom built networks that move the data from one place to another a tiny fraction of a second quicker than other networks. They then capitalize on having that data earlier to eek out small profits at scale.
Imagine two investors sitting at a restaurant table discussing trades they are about to make. The trades they are making will be significant, in the sense that their trades will then impact the value of the stocks they're trading. Meanwhile, a waiter at the restaurant makes a habit of eavesdropping on the conversations of these investors. When he gets the information, he runs to the phone and effects his own, smaller trade.
It's not that the waiter happened to overhear something. The waiter makes it his business to "overhear." The waiter adds no real value. He's a parasite on the people who do add value. The HFT traders are likewise.
What they do is documented in Michael Lewis's Flashboys.
If prices are out of line, that definitely seems like a place where automated trading would be valuable. But HFTs' tech advantage isn't just about finding a better arbitrage algorithm or being smarter: it's also quite clearly about exploiting a pure technical advantage over other traders. The fact that HFTs are so willing to invest in speed even to achieve a tiny advantage over other HFTs kind of gives the game away.
And when people quite rightly observe that having a speed advantage over other traders obviously allows extractive behavior like frontrunning, a bunch of people on HN come out with irate counter-takes that claim HFT are an unalloyed good -- never something nuanced like, "yes HFT could allow for some extractive behavior, but it's counterbalanced by these advantages which I will explain in detail." (And the corollary, which is an explanation of "why a world where all traders have the same speed advantages wouldn't have all the claimed advantages of HFT but be even more efficient.")
I guess it's also worth pointing out that these responses are usually from people who are involved in the HFT industry in some way, and usually they start out by accusing people of "not understanding the industry". Which is precisely the accusation many make against HFT: that it's so deliberately opaque that people outside the industry can't possibly determine how much extraction there is compared to value being added. Saying "trust us" or "you couldn't possibly know because you aren't on the inside profiting from it" is not the compelling argument you think it is.
On the contrary, equities markets need the liquidity HFTs provide to function well. Before HFTs there were human market makers who played a similar role in equities markets, but charged far more to do it. HFT is a story of efficient technology bringing prices down a lot and shrinking the profitability of an existing industry.
I’m sympathetic to the view that a lot of resources have been poured into making trades happen in 1 nanosecond when 100 or 1000 nanoseconds would probably be fine for all practical purposes. The issue is that people keep finding that letting the fastest system get the trade actually does work better than alternative approaches.
What you've said sounds very reasonable to me. I find it interesting that the line of reasoning you have presented doesn't appear to require super-fast trading. That there has been an "arms race" to become faster and faster fits perfectly, because the competition is over who is the fastest market maker. Given that the skill of people/firms engaged in that race has been to be as fast and smart as possible, it also makes sense that they would be less than enthusiastic about any limitation on their competitive abilities. Neither of those points, however, implies that being able to trade as quickly as possible is actually "useful" to the wider market - only that it is a consequence of that market.
It seems possible that there is a lot of effort being put into that endeavour that is, in some sense, wasted - it is spent on competition rather than "production", and with a judicious rule change the cost of competition could be reduced. The inevitable free market counter-argument is that the market will have already selected the best balance of cost/benefit, but I don't find that convincing at all, even if only because obtaining a sufficiently free market is impossible.
Anyway, thanks for your response, it certainly seems (to a layman such as myself at least) to be a good description of the basis for HFT.
It doesn't. These kinds of actions are entirely competition between HFT firms (the fastest ones get the biggest share of the pie). But the HFT firms as a whole do provide a service to the market, and at a much lower price than the alternatives, even with the apparently extravagent cost of these efforts (mainly because it requires a lot less people: old-school market making involved a lot of people on trading floors).
The major issue is HFT can create wild market swings with little to no basis in reality. It's actually possible for them to suck up all the outstanding bids over a few seconds using small amounts of capital and while a human might desire to sell if a stock goes up by 2% the seconds or minutes it takes US to make that choice is eons for the algorithms. The net result of this is actually less liquidity as someone buying or selling can't place large orders on the market or the algorithms with eat them alive. Also, they are often setup to simply stop all actions if the market deviates to far from the norm which pushes things even further out of whack.
That would be totally boring trading and unremarkable, but that isn't what HFT firms are doing. If that were the strategy then exchanges would presumably just sell a skip-the-queue service when two traders offer the same price and there wouldn't be much point worrying about how far machines are from the exchange.
HFT trading isn't about executing the same trade as someone else but a tiny margin faster. That wouldn't have any special impact on market spreads or liquidity, for example.
There aren't any complaints against the T in HFT; as traders they are helpful. The value questions are about the HF and whether it is a desirable part of the market or an unhelpful arbitrage opportunity created only by implementation details of the exchange.
The HFT is taking risk though. Like everything else in markets, nothing they do happens at the same time either. They could miss on one side of the trade if another player is faster, and then they're stuck in a position they might not want.
Also think about this in terms of what happened, the two counterparties got to interact without being in the same place or product. Yes, they could have traded if they were both in the same place, but they weren't. Surely providing that service is worth something. You could buy your meat wholesale, but the grocery store isn't ripping you off by selling it for $1 more a pound somewhere closer to home.
Think about what would happen to others if the HFT wasn't there, too. Prices would diverge and traders on one of the exchanges would be trading away from the real value. By keeping prices in line, information is transmitted from market to market more efficiently. A trader who isn't concerned with cross-market arbitrage doesn't need to connect to and watch multiple markets to get a fair price.
HFTs are obviously motivated by profits, but the level of profit is extremely low for the amount of service they provide. Virtu trades over 5% of the US markets and made slightly over $100 million. They only have a couple hundred employees. Think about how many people were required to make markets globally before technology. Only a fool would argue that we should go back to that system.
As far as I can tell, the commonality among those always-winning firms is high frequency low latency algorithmic trading. These days it takes millions of dollars per month just to pay for the infrastructure you need to be able to be competitive - and then you also have to have some nontrivial edge, without which there isn’t all that much profit in having low latency.
And that is why I like the idea of a "trade arbitrarily slowly with limited price change" market versus the current approach of "trade fast with an arbitrary price change" market.
See https://news.ycombinator.com/item?id=24760841 for an explanation of how the trade arbitrarily slowly market could work. Under normal conditions, it would look a lot like the current market does. Except that you're paying less to the HFT folks.
What I didn't describe there is that you could even have a chain of slower and slower markets. With a maximum rate of price change varying from 1% per day to 1% per minute. With the idea that ordinary folks would trade on the 1% per minute market while large institutional orders would be likely to go in the 1% per day market. (And when the price of two markets cross, open orders on the one can match as open orders on the other.)
Trading is cheap now... because it's all been computerised. The exchanges are digital - as soon as you can do all the transaction processing by computers there is an entire army of (expensive) clerks that can be laid off who used to handle the paper work. HFTs occupy a niche where they can exploit latency between exchanges for profit, with occasional other activities that were considered to be far more questionable when it was people playing those games, not algorithms.
With regards to the mechanics. All I have is what was explained to me by someone who started doing HFT more than 10 years ago and from those traders I know who made money by trading the same stock several time a day and taking a small fee each time they did. Back then they just didn't have computers or the knowledge to do it fast enough I guess.
There is a lot of bot-driven trading going on. To the degree it starts approaching HFT-levels of speed I'm not sure. I don't know what sort of algorithmic trading arms race is going on in that space since the order books are still thin enough for a small number of well-moneyed people to impact the price in a non-trivial manner.
One of the biggest justifications for HFT that I see is that it increases liquidity. However, did major markets ever really have a huge problem with lack of liquidity, 20 or 30 years ago before HFT?
I can't help but wonder if this level of liquidity is only really useful to HFT, if it is something that HFT is both the primary provider and beneficiary of, and if they're sort of using their existence to justify their existence, so to speak.
Chris uses mom and pop trader examples to explain the concepts here, but the fact is that Algo trading now accounts for much of the volume on major exchanges, and HFT is probably a decent portion of that.
I'd love to see Chris or any other HFT trader's take on that. He sort of does near the end, but would be interesting to see a more in depth discussion:
>Although the latency competition provides little to no value to the end consumers, all HFTs must play the game. If they don’t, a faster HFT will beat them to market and their order flow will be reduced.
This is a very real social cost of HFT - many very smart people spend a lot of time and effort reducing the latency of trading systems. I’m no fool, but when I worked as an HFT I often felt like one. The field contains a lot of very smart people, and the world would probably be better off if the stock market had a little more latency and those smart people were building products for the world. For example, after leaving HFT I built a useful consumer product.
But with current market mechanics, the global optimum where trading is marginally slower and smart people build useful products is impossible to reach. The latency arms race has us stuck in a suboptimal Nash Equilibrium which is similar in character to signalling competitions (e.g., the education bubble).
I get that arbitrage helps set the price, but what does HFT do, beyond deciding which of the competing nano-second enabled traders get to edge a profit in an artifact of how information flows?
I don't see this HFT as being information-worthy arbitrage. I have an intent to set a real-world price, I declare it, and in the increment of time it takes my packets to flow, somebody edges in the queue and inserts trades to make me have to deal with them for some increment of price. I wind up paying more, and the original source winds up getting less, or some other event.
its like re-intermediation, in a world heading to dis-intermediation.
What is the real-world, wider economic "good" here? If my pension fund is routinely profiting from this, I am not sure that is a wider good btw: We could probably make more money selling Heroin, if we're going down this 'whatever it takes' track...
The fact they made crazy profits initially is more of an artifact that they were competing against humans rather than other automated systems. In a few years we may see them reach the point of diminishing returns where all exchanges share a common pool of liquidity that adapts in a few millis to new information.
reply