[By Shaunak Rohit Wagle]
The author is a student of Maharashtra National Law University, Mumbai
For more than a decade, the Securities and Exchange Board of India (SEBI) has experimented with ways of taming algorithmic trading. Circulars in 2012 and 2016 addressed risk controls for brokers and exchanges, and a 2025 circular aimed to clarify obligations in the rapidly evolving “retail-algo” space. However, none of these attempts had ever been codified in the SEBI (Stock Brokers) Regulations, 1992. This changed in August 2025, when SEBI proposed to incorporate the following statutory definition of algorithmic trading in the aforementioned regulations:
“Algorithmic Trading” means any order generated/placed using automated execution logic.”
Prima facie, this is a straightforward act of consolidation; however, in substance, it is a radical expansion. It entails the inclusion of every order touched by automation, from the most sophisticated high-frequency strategy to retail SIP auto-executed through an application programming interface (API) in the definition. By making the definition broad instead of precise, SEBI risks blurring vital distinctions, overburdening small intermediaries, and stunting innovation and growth in India’s fintech ecosystem.
SEBI’s draft definition gives rise to doctrinal ambiguities and potential economic burdens that warrant careful reassessment. SEBI should rework its approach through a tiered definitional framework, a retail sandbox, and a clarified liability allocation. A well-balanced framework can help SEBI fulfill its dual statutory mandate under Section 11 of the SEBI Act: to protect investors while promoting market development.
The Problem of Overreach
At its core, financial regulation derives legitimacy from statutory authority. SEBI’s mandate under the SEBI Act, 1992, is straightforward: regulate intermediaries, not clients or software vendors. Yet, by defining algorithmic trading as “any order generated/placed using automated execution logic,” the draft threatens to expand SEBI’s jurisdictional powers indirectly to actors far outside its ambit. While the definition is housed within the Stock Broker Regulations, its practical implications extend further. A stockbroker’s compliance obligations inevitably shape its commercial relationships. If every automated order is deemed ‘algorithmic trading,’ brokers will be compelled to impose stricter due diligence, contractual obligations, and potential liabilities on the fintech firms and vendors that provide API access and other automated tools. This creates a de facto regulatory burden on these entities, as they must conform to the broker’s heightened requirements to remain in business.
This means that a broker using basic order-routing software would, under this definition, be deemed to have engaged in “algorithmic trading”. A client using an API to execute recurring trades might also fall within its scope. This is not because SEBI would regulate the client directly, but because the broker, who is the regulated entity, would be obligated to treat the client’s automated instruction as a regulated ‘algorithmic trade.’ Consequently, the broker would need to subject the client to more rigorous monitoring, risk management protocols, and potentially restrictive terms of service, thereby indirectly bringing the client’s actions under the ambit of the regulation. This interpretative sprawl creates doctrinal instability. Delegated legislation cannot extend beyond the scope of the parent statute. The overreach is not one of direct regulation but of indirect consequence, where the broker acts as a conduit for regulatory burdens that ultimately fall upon their clients and technology partners. Thus, SEBI risks straying into ultra vires territory by sweeping in activity that is not meaningfully broker conduct.
Definitional Ambiguity
The absence of clarity in the definition is also a problem. The phrase “automatic execution logic” is not defined. It raises multiple questions, like – Does it mean any pre-programmed function? Does it require decision-making autonomy, or is mere automation enough? SEBI’s own past practice suggests the former. The 2012 and 2016 circulars specifically distinguished between discretionary algorithmic systems and routine automation. The former involves systems making autonomous decisions on parameters like price or timing (e.g., a VWAP algorithm), whereas the latter simply executes a client’s pre-determined instructions without any independent decision-making (e.g., an automated SIP instruction). The 2025 circular went one step ahead and demarcated retail automation as a distinct phenomenon. The draft, however, collapses these definitions into a single catch-all. The result is doctrinal incoherence: a haphazard definition inconsistent with SEBI’s own regulatory history.
Compliance Burdens and Constitutional Concerns
Definitions have significant ramifications as they affix liability. A broad definition does not simply describe; it mandates who must register, what risk controls must be implemented, what audits may be performed, and what liabilities may be attached. SEBI risks imposing compliance burdens where no systemic risks exist by equating trivial automation with high-frequency trading. This essentially undermines the very proportionality required by Article 14 of the Constitution.
Comparing Approaches: India’s more extensive Definition compared to other countries
The dangers of SEBI’s approach are emphasized by a comparative study of other jurisdictions. The European Union’s MiFID II is instructive. Therein, algorithmic trading is defined narrowly: it only occurs when a computer algorithm automatically determines order parameters such as timing, price, or quantity. Explicit exclusions remove order-routing systems, post-trade processing, and data feeds from scope. The EU’s choice was deliberate as it reflects a clear regulatory philosophy: regulation should only apply when an algorithm substitutes for human discretion in setting economically significant variables.
Similar trends are seen in the United States. The SEC’s Market Access Rule requires brokers to deploy risk controls, but does not attempt to regulate “algorithmic trading” in the abstract. The CFTC’s proposed Regulation AT, ultimately withdrawn after facing significant industry opposition over its high compliance costs and controversial source code repository requirements that raised intellectual property concerns, was designed to apply only to automated systems that generated order parameters, excluding tools used solely for order management. Similarly, the frameworks of Singapore’s MAS and Hong Kong’s SFC define algorithmic trading as systems that make independent trading decisions, while tailoring compliance requirements to match the level of system complexity.
India is perhaps the only country amongst other major jurisdictions to have taken such an anomalous stance by defining algorithmic trading as “any order using automated execution logic”. Such a broad definition entails that nearly all trading in modern electronic markets, given that virtually every order involves some automation, could be swept into its ambit.
Economic and Structural Costs:
With extensive compliance departments and existing systems, large brokers and institutional firms will adapt and absorb the marginal cost of these expanded obligations. However, small and mid-size brokers, many of whom operate with thin margins, will be disproportionately affected. If every API order is “algo-trading”, they will be forced to invest in expensive alternatives such as surveillance systems, audits, and certifications far beyond their risk profile. The inevitable result will be market consolidation, wherein smaller players exit or merge, leaving a more oligopolistic market dominated by a few large incumbents. Investors’ choice and competition within the industry suffer.
A greater risk is borne by the fintech ecosystem. India’s recent surge of retail algo-startups that offer back-testing platforms, retail APIs, and strategy-as-a-service thrives on regulatory space to innovate. Sweeping them into the same regulatory net as high-frequency firms could prove fatal. This creates a risk of regulatory arbitrage, where innovation and capital may be driven towards jurisdictions like Singapore or Dubai, whose welcoming sandbox frameworks stand in sharp contrast to a restrictive domestic environment. Instead of solidifying India’s position as a fintech hub, SEBI risks stifling it.
This chilling effect on the fintech ecosystem is compounded by how overbreadth undermines regulatory efficiency. If every automated order is deemed as “algo-trading”, SEBI’s surveillance will be overwhelmed. Regulators will expend scarce resources monitoring trivial automation rather than genuine systemic threats. This inefficient allocation of regulatory resources not only weakens investor protection but also exacerbates the economic costs imposed on the market, creating a framework that is simultaneously burdensome and ineffective.
Towards a Calibrated Framework:
To remedy the doctrinal, constitutional, and economic issues highlighted above, SEBI should adopt a more proportionate and principled framework that is not only doctrinally defensible but also economically sound. The following are a few reforms that offer the same –
- A tiered definition – The most fundamental reform is to replace the one-size-fits-all definition with a more tiered framework. At the first level lies basic automation, i.e., order routing, SIP auto-execution, and form auto-fill, which should be excluded entirely. At the second level lies algorithmic trading, defined as systems that autonomously set order parameters. These should be subject to risk controls, testing, and audit obligations. At the third level lies high-frequency trading, defined by its sensitivity to latency and co-location, which should be subject to the strictest obligations. This tripartite structure achieves proportionality, which the proposed definition currently lacks. Crucially, the purpose of such a tiered structure is to enable proportionate regulation, ensuring that each category is subject to a distinct level of oversight and compliance obligations commensurate with its systemic risk.
- Regulatory sandbox for retail algos – India’s fintech sector is vibrant because of retail participation. Instead of bringing every retail algo into the full regulatory net, SEBI could permit innovation within a controlled environment. Firms within the sandbox would disclose risks, operate under caps, and report outcomes to SEBI. Investors would receive explicit warnings that sandbox strategies are experimental in nature. SEBI’s option of pioneering a retail algo sandbox is not merely a speculative suggestion. RBI’s fintech sandbox has already demonstrated the feasibility of this approach. Extending this model in the securities market would position India as a global leader, as no major jurisdiction currently offers a sandbox specifically for retail algorithms. SEBI would not only mitigate risk but also showcase regulatory innovation.
- Clarified Liability Allocation – SEBI must clarify the allocation of liability. If a client-deployed algorithm malfunctions, is the broker liable? What about the vendor? Or the client himself? SEBI’s draft offers no answer. A principled framework would allocate responsibility across three parties. While SEBI’s direct regulatory authority remains with the brokers, their compliance function would necessitate a clear allocation of downstream accountability. Brokers would be accountable for enforcing SEBI’s risk controls. Vendors offering certified algorithmic products would be contractually responsible to the broker for meeting compliance standards. Clients, through their agreements with the broker, would be held liable for any manipulative or fraudulent trading strategies. This three-way approach prevents moral hazard, protects brokers from undue burden, and ensures accountability throughout the system.
For instance, Hong Kong’s SFC offers a model: it explicitly recognizes vendor accountability, while maintaining ultimate supervisory responsibility with licensed intermediaries. India can adopt this to its broker-driven markets.
Conclusion
By defining algorithmic trading as “any order generated/placed using automated execution logic”, SEBI risks doctrinal overreach, economic distortion, and regulatory inefficiency. The costs are imminent: chilled innovation, market consolidation, diluted enforcement, and potential constitutional infirmity. The reforms suggested in this article, i.e., a tiered definition, a retail algo sandbox, and clarified liability, together form a principled, proportionate, and innovation-friendly framework. These refinements would align India with global best practices while allowing it to lead in areas, such as retail algo sandboxes, where no jurisdiction has yet ventured.
If SEBI prioritizes precision over breadth, India can safeguard investor interests, foster innovation, and cement its place as a global financial hub. If it chooses otherwise, the costs will be borne not only by brokers and fintechs, but by the very investors SEBI is sworn to protect.