[Pankti Rupani and Arihant Sethia are students of Gujarat National Law University]
Introduction
Artificial Intelligence trading (AI trading) refers to trading with zero human interference. AI trading relies on the utilisation of a predetermined set of instructions in the form of algorithms to generate trading signals and execute orders. AI systems acquire proficiency through exposure to real-world dynamics, optimising task performance to maximise rewards. The programmer only needs to define the payoff metric, with the AI autonomously formulating decisions or procedures to achieve the defined metric.
In this post, the authors underline the concept of efficiency and its implications for the securities market with the increase in the trend of AI Trading in India (Here). The post navigates through an in-depth analysis of AI trading, dissecting its definition, benefits, and the economic ramifications it poses. The essay further deliberates Efficient Capital Markets Hypothesis (ECMH) to lay down grounds for discussion. Finally, the essay proposes a few recommendations aimed at fortifying market efficiency while harnessing the advantages of AI trading.
What is the meaning of Efficiency in Capital Markets?
Efficiency means the optimal use of minimal inputs to obtain maximum output. Capital markets attain efficiency when the asset prices quickly and accurately reflect all available information. Any factors or practices that hinder the market's ability to allocate resources optimally are said to undermine efficiency in capital markets. With the ability to quickly process and act on information, algorithms may exploit temporary price inefficiencies, suggesting that markets may not always be perfectly efficient. While AI can improve market performance in a number of information-related areas such as Robotics, Banking and Financial services etc., it also comes at a cost to the market's capacity to allocate resources productively.
Market efficiency suggests that prices of securities in the market reflect the available information in the market about that security. The ECMH, though debated, asserts that security prices encompass all known data. The hypothesis is based on the notion that the prices of a security represent all the available information in the market regarding that security.
This idea is crucial for investors, policymakers, and regulators alike. It serves as the foundation for various investment strategies, risk management practices, and policy decisions in securities markets. Therefore, regulations in the securities market heavily rely on the concept of efficiency in the capital market. Despite some imperfections in the hypothesis, the interplay between market efficiency, price signals, and capital allocation remains a cornerstone in economic theory and regulatory frameworks. Therefore, determining the efficiency of AI trading can give the best overview regarding the regulations to be formulated for the efficient functioning of the securities market.
How AI Trading Undermines Capital Market Efficiency?
Undermining capital market efficiency refers to the practice of hampering the ability of the market to allocate resources in the most optimal manner. Thus, the efficiency of the securities markets gets undermined when the prices of the securities get far from the equilibrium due to certain factors (Here). This efficiency is usually achieved by means of regulations and mechanisms about price, information, disclosure, etc. This efficiency ensures that capital flows to its most productive uses, benefiting both investors and the economy.
Model Risk
Models are often used in finance to predict future stock values, locate trading opportunities, and assist managers in making business choices. Model risk arises if judgements are made using a model that is not sufficiently accurate. Simplified assumptions about market behaviour are used in models, which frequently overlook the intricacies of human decision-making and changing market dynamics. These mistakes may produce unreliable results, causing algorithms to interpret data incorrectly or forecast market activity incorrectly.
Several such incidents have been seen in the past, one of which was a $20 million loss faced by Samathur Li Kin-Kan. This incident highlights the vulnerability of relying solely on predictive models in financial decision-making. Inaccuracies or flaws in models can have severe consequences, leading to substantial financial losses for individuals and organisations.
When these models, which are central to AI trading, fail to account for nuanced market behaviours, they can distort trading strategies and decision-making. This distortion can further lead to suboptimal market outcomes, hindering the efficient allocation of capital and distorting price signals. Therefore, while models offer valuable insights and aid decision-making, the potential risks and limitations they carry should be carefully considered and mitigated to ensure more robust and accurate financial strategies.
Creativity vs Constraint
The balance of creativity and constraint act as a crucial dilemma for traders in AI trading. Tough programming ensures a pre-determined pattern of action by the tool, but excessive constraints often hamper the adaptability of the market to evolving market behaviours. Studies on flash crashes, i.e. an extremely rapid decline in the price of a stock, underscore how constraints in algorithms give rise to market stress, which ends up as a crisis. There's a debate about whether AI can adequately account for and reflect market risks during crises, potentially raising questions about its ability to trade effectively in unpredictable situations while still ensuring market stability.
Implication for Capital Allocation
AI Trading enhances short-term efficiency as the new incoming data swiftly reflects and enhances the market’s responsiveness. However, it poses challenges for continual fundamental allocative efficiency due to model risks and a potential decline in incentives for investors to engage deeply in research. Thus, high-speed algorithms diminish incentives for fundamental traders to invest in long-term research, affecting their motivation for engaging in governance activities, such as shareholder monitoring.
Further Algorithms have disrupted the traditional efficiency of markets driven by informed traders. These AI-powered entities, especially those engaged in High-Frequency Trading (HFT), employ order anticipation strategies to undermine the profits of well-informed traders. Despite possessing insider knowledge, these anticipators prioritise immediate gains over filling basic informational gaps (Here). Consequently, the reduced commitment of informed traders to governance and monitoring due to lower returns may alter market dynamics. Conversely, this scenario could prompt a more strategic intervention, potentially enhancing the effectiveness of shareholder initiatives.
Proposed regulations
As discussed earlier, regulations in the securities market heavily rely on the concept of efficiency in the capital market. Considering the intricate dynamics that are introduced by AI Trading in the securities market, it comes with utmost importance to make regulations that not only neutralise the inherent risks but also improve the efficiency of the Securities market. Here are a few regulations, proposed by the authors for bolstering efficiency in the securities market:
Enhance model risk governance
Traders can be mandated to establish comprehensive controls around model development, testing, monitoring, and validation. Standardised protocols for model documentation, sensitivity testing, and error reporting can be established. There can be mandatory registration and auditing of proprietary AI tools. Regulatory scrutiny should extend to different algorithms and models to enable authorities to evaluate their intricacies. Regularly publish aggregate metrics on AI behaviour to provide insights to stakeholders and will bring transparency to the securities market.
Mandatory Disclosure Requirements
Mandatory disclosure requirements should be implemented to prime algorithms to react to strong-influence phrases and words that facilitate short-term trading based on relevant information. Mandatory disclosure aims to make the information readily available for the traders. The motive of individual traders to make money from private information drives the markets towards efficiency. The regulation ought to promote the success of these incentives.
Protect informed trading
In order to shield the informed traders from the exploitative practices of HFT, implement batch auctions or minimum order resting times to prevent rapid order anticipation. Implement measures to regulate the speed and flow of trading. The current trend of rewarding the fastest trader should be mitigated. In a more controlled environment, informed traders can be relieved from the pressure to continuously compete on speed and technology.
Conclusion
To sum up, AI trading yields a plethora of benefits but concurrently raises substantial costs on the primary function of securities markets, which is the efficient and effective allocation of capital in the economy. The rise of AI trading presents a nuanced landscape where short-term informational efficiency coexists with potential drawbacks for long-term fundamental allocative efficiency. The rapid information processing and trade execution capabilities of AI have undoubtedly enhanced short-term efficiency; however, this enhancement comes at a price that could compromise the larger goal of allocating funds to promote long-term economic development.
With markets expected to grow more automated, this poses a crucial question for regulators tasked with shaping current and future market structures and laws that govern them. The growing prominence of automated systems in capital markets necessitates a judicious and forward-looking regulatory approach. Measures around governance, transparency, circuit breakers, mandatory disclosure requirements and protecting informed traders can help achieve this balance. Fostering policies that strike the right balance between tapping the benefits of algorithms and ensuring market stability remains imperative. Getting the policy mix right will be the key to harnessing the promise of technology while maintaining market efficiency.
Comments