Unleashing Billions at the Speed Of Light; The Story of High-Frequency Traders
If you could get your information ten microseconds before any others, what would that do for you?
Ordinarily, we would probably shrug and say “Well, nothing.” Anybody who works in the stock market, however, will by now have been frothing at the mouth; some of them, as said by Brennan Carley, would sell their “grandmothers for a microsecond.”
In the world of high-frequency trading (HFT), microseconds matter. A little too much. The obsession with speed has transformed the stock market into a technological battleground, where the fastest and most sophisticated algorithms compete for supremacy. Exchanges such as the NYSE have transformed from arenas of shouting “buy” and “sell” to cold, heavily guarded server rooms all with the intent of creating the most profit.
High Frequency Traders, or HFT firms, are those who manage to send out huge numbers of orders in fractions of a second by essentially front-running the market: pinging little orders of 100 shares or so, in order to be able to see what’s going on. Thus being able to react to market movements before it becomes available to other traders or updated on exchanges.
What enables them to race ahead? Well, one major factor is co-location. In order to be able to compete, HFT firms have to co-locate their trading servers in proximity, usually nearby or within the same building as the exchange’s matching engine; this is why data centers for popular exchanges like the NYSE and Nasdaq have expanded significantly. The matching engine is what essentially matches the buy and sell orders in the market.
Besides this and having colossal technological advantages over ordinary trading systems, HFT firms are able to utilize PFOF or Payment For Order Flow to their advantage (note: this practice is banned in Canada and the UK). This practice means that banks are paid to route their orders through market makers, such as HFT firms. While it benefits brokerage firms that they’re being paid to make business and thus have to pass on less cost to their retail customers, high-frequency traders are able to gain insight into retail order flow and learn how different brokers behave. This entire process is usually kept under wraps, as companies are not required to provide explicit details to retail investors under the current regulations.
Let’s talk about three strategies that HFT firms have historically used:
Rebate Arbitrage: This strategy is one of the simplest; High Frequency Traders purchase stock on an exchange that pays them to be a taker, and then turn around and sell that stock at the exact same price on another exchange that pays them to be a maker.
Slow Market Arbitrage: This strategy is probably the most widespread in the world of HFT, costing just retail investors alone at least $5 billion a year. Like the aforementioned example of “pinging”, this strategy is used when high-frequency traders are able to race ahead and capitalize on different latency rates in stock price change; they see a change in price on one exchange and race ahead to another exchange to buy or sell where it hasn’t updated, taking advantage of their latency speeds.
Dark Pool Arbitrage: This billion-dollar strategy comes into use when HFTs exploit the price differences between the slower, more opaque dark pools and the general markets. For example, a large institution puts an order to purchase shares in a stock at $330 in the Crossfinder dark pool (Credit Suisse’s pool), but the price on the public market for that share drops to $329.50. High-frequency traders purchase the stock on the public exchange and race over to the dark pool, selling them for a higher price and thus causing the institution a loss.
In terms of dark pools, banks such as Credit Suisse have become infamous for their actions v stance on HFT. They came out and advertised as the safe pool, but had been in reality selling private access to HFT firms just as much as anyone else. Really, the only bank truly considered “nice” on Wall Street has to be RBC; and they’ve neither opened their own dark pool nor adopted HFT strategies, which says something. In fact, all four founders of the IEX (Investors Exchange) formerly worked at RBC.
[The Investors Exchange is a market primarily known for its tactics for levelling its playing field. Their major innovation is coiling 38 miles of optical fibre in front of their engine, thus reducing the latency advantage for HFT to nil.]
When looking ahead to the future, there’s no way we can even think about it without talking about artificial intelligence. With the increased integration of AI technologies within HFT systems, trading firms can now take more and more advantage of Natural Language Processing. NLP is a branch of computing that focuses on enabling computers to interpret and generate human language. Using further, more advanced versions of this technology, will allow HFT to take advantage of news and social media posts in real-time to actively make sense of and take advantage of market sentiment. This will help build their event-based trading methodology and assist in further analysis of all government and regulatory publications.
Besides AI, neutrinos have been a marked theoretical technology for the future of speed in HFT. What is the relation between particle physics and the world of trading, you might ask? Well, as I’ve said before, speed is absolutely everything in high-frequency trading. Now, because reducing latency is what these firms strive for, the idea behind neutrinos becomes even more inviting; exploiting their high speed to receive information over long distances quickly. The difficulty with this is, while neutrinos do travel at a speed near that of light, neutrino detectors and beams are not cheap. Because there has been no serious commercial implementation, this kind of project could easily run to close to a billion dollars.
Like it or not, HFT is here to stay; and their impact is undeniable.
