Algorithmic Trading and Retail Investors: Rethinking SEBI’s Regulatory Framework

[Priyam is a student at National Law School of India University.]

The Securities and Exchange Board of India (SEBI) introduced a much awaited circular titled “Safer Participation of Retail Investors in Algorithmic Trading” on 4 February 2025, extending the access to algorithmic trading to retail investors. The proposed regulatory framework has been developed after multiple rounds of deliberations considering consultation papers and public comments. Before venturing into the intricacies of this framework and the deficiencies nurtured by it, it is pertinent to define the scope of what is meant by “algo trading”. 

Algorithmic trading (algo trading) is a mechanized method through which large orders are executed using automated pre-programmed instructions which may be based on the changes in variables such as price, volume, and time. The proliferation of this technology among institutional investors is evidenced by a report published by SEBI in September 2024, which showed that during the fiscal year 2024, algorithmic trading accounted for 97% of foreign investors' and 96% of proprietary traders' profits in futures and options. Since the access of this technology had been limited with respect to retail investors, this circular arrives at an opportune time against the backdrop of growing demands for algo trading to be extended to retail investors. However, this blog intends to point out is the possible shortcomings of the proposed framework when other jurisdictions are looked at.

SEBI’s Proposed Regulatory Framework

The circular’s most talked about clause states that stock brokers would be designated as the principal and the “algo providers” shall act as their agents. This means that brokers have the responsibility to authenticate access to API to the algo providers and handle all complaints of the investors with respect to their agents. However, these algo providers must be empaneled with the exchanges for better oversight. The criteria for such empanelment would be provided by the exchanges themselves.  Each algo order shall be tagged with a unique identifier provided by the Exchange in order to establish audit trail. Another point to note is that the brokers and the algo providers are allowed to have revenue-sharing agreements. To mitigate possible conflicts of interest, the brokers are required to completely disclose all the charges to the client.

The circular also categorizes the algorithms offered by algo providers into two: white-box (transparent logic) and black-box (opaque logic), mandating registration for both. The circular enables “tech-savvy” retail investors to be registered with the exchange if they cross the specified order-per-second thresholds; the exact threshold is to be notified later by the broker's Industry Standards Forum.

A Three-Pronged Critique of the Proposed Framework

It is a welcomed move with SEBI now regulating algo trading utilized by retail investors since this practice was already prevalent among these investors, however, in a sphere of regulatory shadow. A comparative analysis however, leads to the claim of this blog that the circular falls short of its lofty aim of balancing “investor interest as well as integrity of the market” (clause 3). The three possible criticisms of the framework are hereinafter delved into.

First, the allowance of a revenue sharing agreement between the broker and the algo provider coupled with the necessity of the algo provider to be registered with the exchange through the broker, may lead to a situation wherein it is in the corporate interest of the broker to prioritize said revenue-sharing arrangements over investor protection. While the circular does state that “the broker shall ensure that such arrangements do not result in any conflict of interest”, on the face of it, this is not a sufficient safeguard since no standards are laid out to determine whether such conflict exists and no possible sanctions are outlined in case the broker fails to do so.  While it would be ideal if such agreements were not permitted to exist at all, if it were the case, the brokers would have no incentive to be designated as the agents of the algo providers. Therefore,  a possible recommendation to the SEBI is to adopt the provisions relating to conflict of interest as outlined in Article 23 of EU’s Markets in Financial Instruments Directive (MiFID II). The article, much like SEBI’s guidelines, requires the firms to prevent or manage possible conflicts of interests between themselves and their employees or agents as the case may be. However, there are a few reasons why SEBI’s guidelines fall short of its counterpart. The major difference between SEBI’s approach and the MiFID II is that while SEBI requires the brokers to disclose the share of charges to the clients, the MiFID II goes a step forward and requires the “investment firms” to clearly and explicitly disclose to the client the general nature and/or sources of conflicts of interest and the ways that possible conflicts have been mitigated. Notably, this disclosure must be tailored to the nature of the client’s profile, enabling them to take an informed decision. This sort of an individualized mechanism is again observed in the following point.

The US deploys a general obligation upon the firms of acting in the “best interest” of the retail customer. This means that the broker shall not place its (or its agent’s) financial interest ahead of the retail customer’s interests. This process involves a full and fair disclosure and elimination of any possible conflict of interest that may not be reasonably compliant with the Regulation Best Interest, 2019.

Second, the circular does not require the congruence of algos and the risk appetite of the individual retail investor. It must be remembered that in a dynamic and emerging market like India, millions of India’s youth have entered the stock market in hopes of amassing a huge fortune. This has led to a culture of these novice investors undertaking large risks in order to achieve their perhaps unrealistic goals. In this backdrop, the SEBI circular does little to ensure that these investors are not swayed by unrealistic returns regularly promised by algo providers. To merely designate the brokers as the principle and algo providers as their agents is a step in ensuring post-trade accountability. However, a better mechanism to further strengthen the regulatory framework is introducing pre-trade “suitability assessments” as seen in Article 25 of MiFID II. These suitability assessments analyze the retail investor’s investment objectives, risk tolerance, and their knowledge and experience to align algorithmic trading strategies with investor risk tolerance. A similar mechanism is found in the UK’s Financial Conduct Authority’s Conduct of Business Sourcebook (COBS). COBS 10.2 places a responsibility on the firms offering these services to assess whether such service would be “appropriate” for the client (based on the client’s knowledge and experience). Without such suitability or appropriateness assessments, retail investors may deploy complex strategies without demonstrating comprehension of the possible risk associated with the same. Introducing these pre-trade mechanisms would be much better suited in their goal of balancing investor interest as well as the integrity of the market. 

Third, SEBI does recognize black-box algos and requires them to be registered as a “registered analyst” and maintain a report for each algo and ensure to the exchange that they have kept such a report. However, this imposes an asymmetrical burden on brokers to be held accountable for such black-box algos. The burden is asymmetrical because brokers lack the technical capacity to audit proprietary algorithms when the logic of these algorithms are not provided. Comparable international jurisdictions have evolved mechanisms that balance innovation and accountability. The EU, for example, through Article 17 of the MiFID II, requires the algo providers to maintain strategy descriptions for all algos, including the details of trading parameters and limits, ensuring brokers can audit general methodologies without accessing proprietary code.

Conclusion

While the concerns with the framework as outlined before are warranted and necessary for creating a regulatory framework comparable with the protections offered in other jurisdictions, it must also be noted that the circular is a huge step forward. Considering the fact that before these guidelines, the entire algo trading market for retail investors thrived in an unregulated sphere, one must appreciate this circular for its timely intervention in this sector. Further, several provisions put forth are on par with foreign jurisdictions, including the provision of an automatic “kill switch” which halts trading activity based on pre-defined criteria. This is comparable to the provision of circuit breakers as posited in paragraph 5 of Article 48 of MiFID II and the US Commodity Futures Trading Commission’s 2015 proposed rules. 

There is still room for the authorities to include or change certain provisions since the exchanges have time till 1 August 2025 to amend the laws and provide guidelines in tune with this circular. Certain changes that would be surely beneficial include necessary suitability assessments, broad descriptions of black-box algorithms, and those necessary to mitigate possible conflicts of interests between the brokers and the algo providers.