Home > Articles > Ethical Implications of High Frequency Trading Algorithms

Ethical Implications of High Frequency Trading Algorithms

The modern era of finance has brought about a multitude of technological advancements, from simple online brokerages to the hoard of available information on the internet. Ultimately, the ubiquitous nature of technology within finance has led to newfound responsibilities for software engineers. Programmers employed by large firms have increasingly begun to leverage technology in their algorithms. As a result, average trade size has fallen substantially as programmers enact ”high frequency” trading strategies. These strategies, often deemed ”high frequency trading” or ”HFT” for short, have altered the way markets work around the world. They also raise the question of whether it is ethically permissible for software engineers to enact HFT algorithms.

In this paper, we provide an answer to that question using consequentialism, the ACM code of ethics, and evidence from the industry. Programmers using HFT algorithms can be ethically permissible due to increased market liquidity and efficiency, reduced costs, and profitability. Alternatively, detractors consider an engineer's role in HFT not ethically permissible because of potential market manipulation alongside the cascading effect HFT can have on the markets. To reach a conclusion, we analyze the theory of consequentialism, view consequences of programmers coding HFT algorithms, and provide ethical support for their continuance. Lastly, the legal ramification of no longer needing to advocate against HFT will be discussed before offering concluding remarks. Taking both sides into consideration, this paper postulates that programmers engaging in HFT is ethically permissible because it fulfills both the optimific framework of consequentialism and multiple ACM ethical guidelines.

The Ethics of HFT

High frequency trading describes the combination of mathe- matical modeling and computer science algorithms. An example of an algorithm commonly developed is statistical arbitrage. The method compares historical data from stocks that typically move up or down simultaneously. Programmers use algorithms that track their movements and when one stock drifts apart, automated trades are executed to take advantage of the price difference. Highly efficient time complexity solutions are a necessity, so programmers hired possess a thorough command of C++ or Python.

Consequentialism states that ”an action is morally required just because it produces the best overall results,” which is synonymous with an action being optimific. More specifically, we use “act utilitarianism” which acts as a subset of consequentialism. In section four, the specific steps toward determining whether programmers writing HFT code qualifies as optimific will be examined.

In practice, programmers are given the following guidelines: trading regulations with which their algorithms must comply, the firm’s code of conduct that states one should serve the best interests of clients, and the assurance that they provide liquidity to the markets. These rules are difficult to enforce, though, considering the highly secretive nature of hedge fund strategies. This can lead to certain problems, both morally and economically, which we analyze in the next section.

Advantages and Externalities

HFT has several potential and realized ethical pitfalls. Consider a first example, involving the NYC firm Trillium Capital. In 2010, Trillium programmers used code to influence the market in their favor. They commenced a series of trades that are referred to as “spoofing,” which occurs when trades are entered without the intention of following through on them. This then induces others to initiate trades based on the illusion of demand. Trillium profited from the increased demand before selling their holdings. Trades of this caliber are illegal because they cause market movements that would not ordinarily happen. The scheme is also sometimes deemed a “pump and dump” due to the artificial pumping up of the security before it is subsequently dumped.

Additionally, the “flash crash” of 2010 is another event largely attributed to HFT. When algorithms across firms each share similar criteria for execution, they can create a cascading effect across the entirety of the market. This can cause instantaneous crashes such as the one observed in 2010.

On the opposite end of the spectrum, we see merits of HFT. The SEC concluded that HFT improves market liquidity. Improved liquidity leads to lower prices for retail traders. At the same time, stakeholders receive more benefits from HFT than negatives. The economy, financial system, political system, consumers, and individuals and their families are all identified as positively impacted stakeholders. HFT even promotes ethics by forcing competition between other financial companies within sectors that are perceived to be less ethical. Each of these positive aspects work in favor of programmers coding HFT algorithms.

Thoughts on the Current Discourse

To begin our application of consequentialism, consider the following steps that determine whether programmers utilizing HFT are acting optimifically:

  1. To begin, it is necessary to identify what is intrinsically good. Common selections include happiness, autonomy, knowledge, and virtue.
  2. Next, consider the opposite–what is intrinsically bad? Mental and physical pain, betrayal of innocents, and sadistic tendencies are cited as meeting this criteria.
  3. Programmers engaging in HFT have a wide variety of options when creating algorithms. They can pursue maximal financial gain by exploiting regulations. They can increase knowledge of the financial mathematics sector of academia. They can put their own happiness first and prioritize work-life balance. This encompasses the third step; what options do the programmers have?
  4. Assigning each of step three’s actions to a value is difficult, but not entirely incomprehensible. Prioritizing financial gain by breaking rules falls closer to “bad,” while increasing happiness and knowledge are both good options. This leads us to the final step.
  5. Step five says that we need to make the optimific choice. Following our criteria in steps one and two, it is evident that maximizing knowledge and the happiness of consumers would lead to the most good in the most people.

That was a thought exercise that demonstrates the optimific principle of consequentialism. But what occurs in reality? If reality follows the optimific choice closely, then it will be safe to conclude that programmers engaging in HFT are behaving ethically.

Recent years have shown HFT to be increasingly interesting in academia, as academics have begun researching market microstructure. There are around 3,000 academic papers on academia.edu related to the topic, which speaks to the knowl- edge incited by HFT. Software engineers inherently respect the ideas of their colleagues by using citations, which follows ACM code principle 1.5. HFT also improves the happiness and financial well-being of consumers, which again stems from the ACM code. It is estimated that programmers utilizing HFT save Vanguard investors hundreds of millions of dollars each year. Executing more trades at a faster rate increases market speed while lowering overhead costs.

Market manipulation tactics perpetually exist, and they re- main immoral. Programmers engaging in HFT as it is intended avoid harm, as stated within the code of ethics. This appeals to the concept of expected results within consequentialism. Software engineers are not attempting to perform algorithms that lead to “flash crashes” and manipulation. Anyone in their shoes can reasonably expect to increase liquidity and profit for investors.

Overarching Implications

Since we have determined that programmers formulating HFT algorithms is ethically permissible, that means there are implications stemming from that realization. One which we focus on here is legality. In 2013, the SEC introduced MIDAS to regulate cases of spoofing. With our conclusion in tow, MIDAS remains relevant to reduce exploitation. The only legal implication that would alter would be to eliminate the stance that HFT should be eliminated. As we alluded to above, ridding the markets of HFT would reduce liquidity, thus raising prices for all consumers. Since programming HFT algorithms is ethically permissible, the legal onus to abolish the practice is no longer necessary.

Concluding Remarks

Technology has forever changed the financial industry. Questioning whether current practices are unethical will ensure individuals are held to the same standards as they previously were, regardless of technological advances.

In this paper, we have applied this line of questioning to software engineers utilizing high frequency trading. Using principles of consequentialism, it is evident that programmers designing HFT algorithms are not behaving unethically. Their goals include improving knowledge and happiness, while any negative consequences can be reduced to unanticipated events. Overall, the results are optimific and yield the best ratio of good to bad, which allows us to conclude that programmers coding HFT algorithms is ethically permissible.


Sources