by I.R. Mullin
Copyright©2014 I.R. Mullin. All Rights Reserved. No part of this book may be reproduced or retransmitted in any form or by any means without the written permission of the author.
Should you have any questions, please contact us at firstname.lastname@example.org
Part I of this book familiarizes readers with the nature of high-frequency trading (HFT) and the issues created by HFT.
I.1. How Does High-Frequency Trading Work?
At the end of World War II, investors held a typical company’s stock for four years. By the year 2011, this time had decreased to twenty-two seconds, and now the fastest computers process quotes in a millionth of a fraction of a second. The modern market looks like a wild place supercharged by advanced technologies. And it is still picking up speed while it changes the rules of the game. Trading does not take place on a stock market floor anymore. Instead, it happens inside black boxes located in highly secretive data farms.
Large institutional investors, such as banks, mutual funds, pension funds, hedge funds and insurance companies, have their orders executed on the market automatically by these special computer algorithms. This is called algorithmic trading. It accounts for more than seventy percent of U.S. equity volume.
Algorithms are mathematical functions that are run by super-fast computers. They calculate where and what to buy and sell, and then they execute the trades automatically.
High-frequency trading (HFT) is one of the types of algorithmic and computerized trading. Sometimes, HFT is also called “black box trading.” It uses technology to obtain information and act on it before anybody else on the market does. Super-fast computers of high-frequency trading firms (HFT firms) divide trading periods into extremely small increments (measured in milliseconds, microseconds, and even smaller increments). They analyze the market, seeking opportunities to take advantage of it, and use computer algorithms, which are also known as “algos” or “high-frequency traders (HFTs),” to front-run other investors.
HFT firms employ physicists, mathematicians, computer scientists and electronic engineers to design million-dollar HFTs. In the HFT industry, these people are sometimes called “quants.”
Each trade generated by HFTs creates a profit in the amount of a penny or even less than a penny per share. This may seem like a very small amount of money, but it easily turns into huge profits when millions of such trades are generated every day. HFTs dominate the market and keep it beyond human control. They can easily front-run mutual funds and pension funds. People who create HFTs sometimes refer to mutual funds and pension funds as “whales” or “dinner.”
I.2. High-frequency trading and Latency Arbitrage
HFT firms profit from being the fastest, and they take advantage of slower traders by being able to track or anticipate them. They do not care what they trade. They simply make money using latency arbitrage. Latency arbitrage refers to the practice of buying or selling shares and other financial products a little ahead of other investors.
A few advantages exploited by HFT firms in order to make latency arbitrage possible include:
- The Raw Feed Advantage—HFTs have an access to raw exchange (direct) data feeds, which are different from the consolidated quote feeds used by the majority of investors. In the U.S., the majority of professional investors and fund managers obtains market data, such as the latest traded prices and best bids/offers, from a consolidated data feed.
The consolidated data feed is relatively slow; it provides investors with the data that is a few milliseconds out of date. HFT firms get the price information and other market data directly from the exchanges. By using this advantage, HFTs can see and trade on prices ahead of the rest of the market.
- The Co-Location Advantage—The shorter the distance that an electronic message has to travel from one point to another, the less time it takes the message to travel. For data travelling at the speed of light, for example, it takes 1 microsecond (one-millionths of a second) to travel 200 meters, and it takes about 25 microseconds to travel 5 kilometers, which is an equivalent of a trip from New Jersey to Manhattan. Latency is determined by the travel time of the electronic message, and it exists for electronic messages sent over greater distances. When a computer processes trading information within thousandths or millionths of seconds, the time it takes data to travel from the exchange’s data center to that computer becomes an important factor.
HFT firms have their trading servers placed (co-located) at the exchanges’ data centers. In this way, they significantly reduce the distance between their own computers and the exchanges’ computer servers.
- Algorithmic Processing Advantage—HFT firms apply sophisticated algorithmic trading that allows them to predict mathematically what trades will happen next. HFTs take a large number of data points and use these data points to calculate where and what to buy and sell, and then they buy and sell as quickly as they can. Each trade generated by HFTs creates a small amount of profit. Millions of such trades generated each day bring huge profits.
- Hardware Advantage—HFT firms use various advanced technologies and methods, such as cut-through switches that can handle data better than conventional Ethernet switches, superior computing processors, etc.
These advantages enable HFTs to:
- predict the trades that are about to happen;
- know prices a few milliseconds before anybody else does;
- jump in front of trades from slower investors and buy shares and other financial products that are readily available on the market, and then, a fraction of a second or seconds later, sell them to the slower investors at a higher price;
- generate a huge profit via millions of small guaranteed wins every day.
Latency arbitrage allows high-frequency trading firms to have a steady stream of profits, with practically no risk involved.
In gambling casinos, the house always wins. Similarly, HFT firms, who own the front of the line on the market, are the house that always wins, and this method also makes the market “rigged.”
I.3. Exchanges and Their Affair with HFTs
Many experts and investors argue that HFT firms have an unfair advantage over other investors and that such advantages create a conflict of interest. The conflict of interest arises because the exchanges have to protect equally the interests of all investors, yet, in reality, the exchanges favor HFT firms over non-HFT investors.
In the past, different U.S. exchanges were offering the same stocks at different prices. In 1975, the U.S. Congress mandated the creation of the SIP (Securities Information Processor) in order to provide investors and professional traders with access to real-time price information for each exchange’s best quotes (bids and offers).
For the SIP, the data about best quotes and trades is collected from all the U.S. exchanges, then the data is processed and sent to the users as one stream of information. The SIP is essentially the central, consolidated live stream and aggregator of every exchange’s best quotes and trades.
The idea behind the SIP was to establish a National Market System that would allow traders to access real-time price information. However, HFT firms have been able to apply different strategies in order to get trading information before anyone else does. As we described earlier, HFT firms compete in placing their servers as close as possible to the floor of the stock exchange. Many experts argue that such practices violate the principle of fair play, since U.S. stock exchanges promise to provide real-time information via their standard SIP feed to the users who pay for this service.
According to the law, the U.S. exchanges are prohibited from sending quotes and trades to direct feeds before sending the same information to the SIP. Nevertheless, in 2012, the NYSE was fined when it was caught sending real market information (presumably because of some unforeseen software issue) to private data feeds before sending it to the SIP, giving HFT firms unfair advantages via the private data feeds.
HFT advocates argue that since 1975, the Securities and Exchange Commission (SEC) has been quite clear that the law requires that data has to be distributed (i.e., released) from the exchange at the same time for all end-users, but that the law does not require that data arrive to all end-users at the same time. This is where co-location becomes so advantageous: exchanges allow HFT firms to have their trading servers placed (co-located) at the exchanges’ data centers, extremely close to the exchanges’ servers.
Because of the reduced distance that an electronic message has to travel, exchanges’ data reaches HFT firms before it reaches anybody else, thus giving HFT firms their time advantage. Non-HFT investors receive the market data a few hundred milliseconds after HFT firms see it. From the point of view of HFT, this amount of time is rather substantial.
In his interview to the Chicago Tribune regarding the inability to “see” the real market, Joe Saluzzi, a co-head of trading at Themis Trading, said, “This is a great example of the two-tiered market that we have developed. There are the ‘haves,’ who subscribe to all of the private data feeds and various hardware and software and spend a ton of money, and then there are the ‘have-nots,’ which are the public who rely on the SIP.”
Today, the exchanges deliver information faster to those who pay for expedited delivery, and they also give such “subscribers” the ability to jump in front of all other investors. The arrangement between exchanges and HFT forms becomes even more complex, though. In his book Dark Pools: The Rise of the Machine Traders and the Rigging of the U.S. Stock Market, Scott Patterson points out that the exchanges make money when trades are executed on their particular exchange, and they have HFT firms utilizing HFTs that get order flow from other traders to a particular exchange.
I.4. High-Frequency Trading and Its Impact on the Market: Good or Bad?
For retail investors (i.e., individual investors who trade in small stocks for their own personal accounts), HFT does not look like a threat because it presumably does not deal with low volume stocks. However, the majority of middle-class investors have their investments in mutual funds and pension funds, and their retirement savings depend on the performance of these funds.
HFTs attack retirement savings of regular investors every time they target mutual funds and pension funds. So, it looks like the finance industry has invented another technique to swindle middle-class investors. Many retail investors (i.e., individual investors) can be disappointed in investing in the stock market because HFTs give a few participants an unfair advantage.
In 2012, Andrei Kirilenko, the Chief Economist at Commodity Futures Trading Commission (CFTC), Jonathan Brogaard, a Finance Professor at University of Washington, and Matthew Baron, a graduate student at Princeton University, reported the results of their study that focused on S&P 500 e-mini contracts.
According to their study, HFT firms earned their profits at the expense of other traders, and this fact might encourage traders to leave the futures market. The study did not cover the equity markets where high-frequency trading makes up a hefty amount of stock trading volume (probably more than seventy percent). However, the authors of the study suggest that the same conclusions can apply to the equity markets.
Moreover, the researchers point out that the race for speed on Wall Street results in profligate spending on human capital and technology. This is because the HFT race for speed forces non-HFT investors, such as traditional trading firms, to employ highly educated people and spend millions of dollars in order to acquire technology that would allow them to keep up with high-speed competitors.
Some experts compare latency arbitrage used by HFTs to a faster version of “SOES Bandits” that the SEC banned in the late 1990s. At that time, the exchanges were dependent on “market makers” to broadcast changes in prices. Market makers were updating their prices at different speeds. The SOES Bandits could pinpoint a slow market maker and complete transactions based on the more up-to-date price information they received from a faster market maker.
One of the problems created by HFTs is called quote staffing. Quote stuffing refers to a tactic of quickly entering and withdrawing large orders. Only firms with a direct link to the exchange can execute these orders. The purpose of quote staffing is to overflow the market with quotes that competing traders and HFTs have to process, thus making traders lose their time and competitive edge.
Quote stuffing creates such a huge volume of information flow that it makes it more expensive for other market participants to process it. For quote stuffing to happen, HFTs create orders that are not real. They are decoys. Placing decoy orders without any real intent to execute a trade is a deceptive operation that may also be considered fraudulent.
Broker-dealers often have to pay a trade execution fee, which is then used by trading venues to pay HFTs a rebate for providing liquidity.
In order to avoid paying these fees, brokers have created dark pools, trading platforms originally designed to anonymously trade large block orders. As a result, trading now takes place across many exchanges and dark pools. This complexity, combined with technological advancements, creates profit opportunities for HFT firms.
Various studies are being conducted to determine the role of HFT in the market. While some studies find that HFT is detrimental to markets, other studies insist that HFT can be beneficial. One should keep in mind that outcomes of the studies may depend on who is paying for the study, and how the questions are formulated.
In 2013, Charles Jones, an economics professor from Columbia University Business School, published a research paper “What Do We Know About High-Frequency Trading?” and an article “The Reality of High Frequency Trading,” wherein he writes, “HFT and related technologies are making markets better, not destabilizing them.” Whereas Professor Jones argued in his publications that HFT should not be subjected to substantial regulations, his research of HFT’s impact on the market was financially supported by a grant from Citadel LLC, one of the largest HFT firms.
HFT advocates claim that HFT provides liquidity and adds competition to the market, driving down the price of stocks and helping investors cut transaction costs. Generally, liquidity refers to a measure of the readiness of the market participants to sell or buy a particular asset (for example, a company stock). The greater the liquidity in a stock, the more likely that an investor will be capable of selling or buying the stock, whenever he wants to do that.
Liquidity is an important property of a well-functioning market. When there is liquidity in the market, a trader can buy or sell without affecting the asset’s price in any significant way. Lack of liquidity creates conditions for a financial crisis. Today, liquidity is often used to measure the success of exchanges, and exchanges pay rebates to companies that add liquidity to the market and charge traders who take liquidity away. In other words, exchanges utilize the “maker-taker” pricing model.
For example, the NYSE’s rebate for providing liquidity is about $.0015 per share. HFT firms earn a profit from this rebate by completing millions of transactions every day. Some trading venues try to attract liquidity providers by offering them information services and super-fast connectivity for a fee.
HFT firms and their advocates claim that HFTs have taken over the role of the market makers (the intermediary), providing liquidity to the market, but this claim is highly questionable. Traditional market makers were buying and selling stocks continuously during the trading day. At any moment in time, there could be a seller on the market, but no buyers. The market makers would buy the shares from the seller, hold them temporarily and then sell to the buyers when they came to the market. The market makers were taking risks when they were buying and holding shares because they did not know when exactly there would be a buyer willing to buy those shares from them.
However, HFT firms are not willing to buy and sell just any stock. They do not take any risks. HFTs buy certain financial products only when they have information that a large institutional investor is just about to buy those financial products.
Once they have this information, HFTs front-run the large institutional investor and buy the financial products only to sell them to that institutional investor a few fractions of a second (or a few seconds) later for a marginally higher price.
If HFTs did not exist, the institutional investor would have executed the trade anyway, because those financial products have already been readily available on the market. Yet, HFT firms receive liquidity rebates from exchanges, although their orders are merely “flashed” orders that are front-running already existing large order blocks.
We should mention that front-running is an old cheating technique. It existed before the emergence of high-frequency trading. In the past, a trading firm, would receive a large order from a client, and the firm would buy some of the same stock for itself before executing the client’s order.
However, just because front-running existed before high-frequency trading, that does not make it the right thing to do. Besides, old-fashioned front-running looks insignificant in comparison to electronic front-running, which is technologically advanced and produces many billions dollars in profits.
Exchanges offer rebates to firms that increase trading volume. Even when an HFT firm buys a stock for $20.00 per share and then sells the stock for $20.00 per share, it receives a fraction-of-a-cent-per-share rebate. When the HFT firm buys and sells millions of shares every day, the rebate brings it huge profits. It is a well-known fact that one cannot get something out of nothing. Rebates paid to HFT firms increase the cost of trade execution to investors who buy the stock in order to hold it for more than a fraction of a second.
Another issue related to HFT is the stability of the market. HFTs’ activity draws the attention of regulators because HFTs can make the market intrinsically much more unstable and predisposed to unexpected and baffling crashes. This serious issue remains unresolved because the magnitude and technological advancement of high-frequency trading make it impossible for the SEC to monitor or regulate the market.
One real-world example of instability created by HFT occurred when Knight Capital, a financial firm, lost $457.6 million in less than an hour. This happened in 2012 when one of its programs, instead of being deactivated, kept placing trade orders for about forty-five minutes, and in each of those extra minutes the firm lost an additional ten million dollars.
The industry had already produced a few big shocks because of high-frequency trading. One of these shocks was the famous “Flash Crash” of May 2010. In the Flash Crash, the market plunged so far that one trillion dollars was wiped off the value of markets in the course of ten minutes. For a short time, some large blue chip stocks traded at one penny, and then just as mysteriously the market rebounded in about thirty minutes.
After all, the Dow Jones Industrial Average closed about four percent lower on the day of the Flash Crash. In its report, issued on October 1, 2010, the SEC announced that a single massive sell order executed by one firm’s algorithm was responsible for the crash. According to the SEC, this trade created a cascading selling effect among HFTs, as well as other traders, and sent the market into a temporary panic.
The Flash Crash made many investors, as well as regulators, consider the necessity of tighter regulation on HFTs. This concern has been supported by the fact that smaller, less noticeable flash crashes keep happening on a regular basis. The original Flash Crash prompted the SEC to impose circuit breakers that would supposedly address this issue.
Circuit breakers, which are also called “collars,” refer to any of the methods used by stock exchanges during times of large sell-offs to prevent panic selling. Quickly falling prices might magnify panic among investors, and circuit breakers are used to provide a cool-down period. Circuit breakers are typically activated when prices cross some preliminary established boundaries, and they cause markets to stop trading for a specific period. Circuit breakers also help during sudden market movements by letting orders accumulate and by preventing the execution of erroneous orders.
High-frequency trading may be changing the nature of the market. The original idea of the stock market was to promote capital formation and flow and to allow individual investors to finance business enterprises. Today, HFT accounts for up to seventy percent of all trades by some estimates, and HFT firms have no interest in the fates of companies themselves; they do not take into consideration company fundamentals (financial statements, business plans and performance). HFTs’ objective is to use superior speed in order to game the stock market and to steal from other investors.
Another concern about HFTs is related to the fact that HFT firms keep their HFTs very secretive. Their secrecy may create conditions for an algorithmic terrorist attack, where a terrorist organization could potentially design HFTs that would cause a market crash and destroy investments and retirement savings.
I.5. High-Frequency Trading and Legislators
So far, regulators have not been able to keep up with HFTs. In 2011, Gregg Berman, the Associate Director of the Office of Analytics and Research in the Division of Trading and Markets at the SEC, appeared in the documentary film “Money and Speed” made by the Dutch filmmaker Marije Meerman. In the documentary, Berman said that there is no one person or group that monitors the market in real time. He also said that “air traffic control” for the market is impossible because it would cost the taxpayers too much money.
Moreover, politicians and lobbyists have created new legislation for stock exchanges and powerful traders that make many of their unethical practices look legitimate. For example, HFT firms organized a Proprietary Trading Group Lobby to act on their behalf in Washington, DC.
In the New York Times’ article “Fast Traders, in Spotlight, Battle Rules,” published in June 2011, Graham Bowley writes, “The group comprises 31 firms. According to data calculated in June, its biggest members spent $690,000 on lobbying last year, more than double what they spent in 2009. They gave more than $547,000 to lawmakers’ political campaigns in 2010, on top of the $456,000 they handed out in the last political cycle in 2008.”
HFT firms also use the fact that policymakers in the U.S. do take into consideration findings of independent academic research when they develop their policies. HFT firms provide various grants to support academic research that presents HFT in a positive light. Earlier in this book, we mentioned the study of HFT conducted by Charles M. Jones, a professor from Columbia University.
In his publications, Jones praises academic studies that support HFT and he practically dismisses studies opposing it. Yet, one of the press releases by Columbia Business School informs us that his research was supported by a grant from Citadel LLC, one of the largest HFT firms.
In 2010, Knight Capital, a financial firm, paid for an academic study that supported HFT, and the CEO of Knight Capital referred to that study in his congressional testimony. Soon after his testimony, Knight Capital lost about 460 million dollars because of a glitch in its computerized high-speed trading system.