Algo Trading: Performance Validation Team May Be Supervised By Credit Rating Agencies
Discussions are going on to determine which organisation would be better suited for the overseeing. Until a decision is made, regulations governing algo trading could remain in limbo.
Credit rating agencies are being considered to oversee a new performance validation agency that will verify claims made by algorithm-based trading platforms, according to people familiar with the matter. Another contender for this role is the NSE subsidiary, NSE Administration and Supervision.
Discussions are going on to determine which organisation would be better suited for the overseeing, and until a decision is made, regulations governing algo trading could remain in limbo.
Proposed Framework For The Agency
A report from the Algorithmic Trading Sub-Group, submitted to the technical group discussing the formation of the PVA, outlines a comprehensive framework for its operations, and has been reviewed by NDTV Profit.
The proposed framework emphasises the importance of registration, requiring that algo trading strategies be registered with the stock exchange prior to validation. Algorithm providers must showcase the performance of their strategies to ensure transparency and accountability.
The validation process is set to be rigorous, with algorithms needing to run in a controlled environment for a minimum of six months. During this period, performance metrics will be calculated according to guidelines established by the PVA. The agency will extract necessary data for validation from market infrastructure institutions in a specified format, ensuring that the information is accurate and relevant.
Once the data is collected, the PVA will verify the performance metrics calculated by algorithm providers. This verification step is crucial to ensure the integrity of the validation process. If the PVA successfully validates an algorithm's performance, the provider will be allowed to use the algorithm for retail trading, as well as showcase its performance publicly.
The PVA will categorise algorithms into specific types, such as arbitrage, execution, or directional strategies, and will determine whether certain types of algorithms, particularly those executing only one kind of trade, can be validated. This classification will help streamline the validation process and ensure that each algorithm meets the necessary criteria.
The report highlights the risks associated with algorithmic trading, particularly the possibility of overfitting models to historical data. This can result in algorithms performing well in backtests but poorly in real-time trading scenarios. To mitigate this risk, the sub-group stresses the need for a thorough and well-structured validation methodology.
Once the PVA is established and fully operational, it is expected to play a critical role in enhancing the credibility and effectiveness of algorithm-based trading platforms in the financial markets. By ensuring that these platforms are held to high performance standards, the PVA aims to protect investors and maintain the integrity of the trading environment.
A mail sent to the Securities and Exchange Board of India as well as the NSE with queries on the matter did not elicit a response.
