S.P. Kothari of SEC Speaks on Policy Challenges Stemming from Big Data

According to Mr. Kothari, the following are the key policy challenges that SEC is facing from big data:

• Security. The volume, velocity, and variety of big data make security particularly challenging for several reasons. This is because big data are not only harder to store and maintain but big data are also bigger targets for bad actors. Therefore, SEC must be mindful of the data it collects and its sensitive nature. SEC continues to look into whether it can reduce the data it collects or reduce its sensitivity. One example of this is SEC recently modified the submission deadlines for Form N-PORT, to reduce the volume of sensitive information held by SEC. This simple change reduced the cyber risk profile of SEC without affecting the timing or quantity of information that is made available to the public.
• Technology. He mentioned the technology arms race between trading firms and the media reports about the increasing use of artificial intelligence, machine learning, and related tools. However, deployment of these technologies involve fixed costs that exclude small, fragmented, or less resourceful market players. Additionally, there are cultural differences between organizations that affect not just the choice of which technology to deploy but also the timing of deployment. For example, hedge funds might be able to adopt new technologies such as cloud computing more quickly than pension funds are able to do so. He also pointed out that certain technologies are "inherently challenging for SEC to monitor." To mention just one example, suppose that artificially intelligent algorithmic trading eventually starts spoofing without the knowledge of the algo creator, (Spoofing is a prohibited activity that involves creating and cancelling a large number of trades in an attempt to convey false information about market demand.) how would SEC respond to this development.
• Communication. A perennial challenge for SEC is to find cost-effective ways to reduce the variety of financial data without loss of substantive information. For instance, SEC has required filers to tag some data using methods such as XML, FIX, FpML, XBRL, and, more recently, Inline XBRL. By dramatically reducing the variety of the data, tagging transitions an electronic document from being human readable into one that is also machine readable. An additional feature of data tagging is network effects. It is well-known that data in tagged 10-Ks can be linked to data from other forms and other firms. Perhaps it is less appreciated that data in tagged documents could be linked across regulatory boundaries and even national boundaries provided the regulator community required similar data tagging. For the SEC, a key benefit of cross-regulator consistency in tagged data is the ability to understand better the nature of the risks in the financial markets. The markets today do not stop at national borders, so looking only at intra-national data provides only a partial picture of the system’s risk.

Out of the volume, velocity, and variety, perhaps the best way to make future data sets more manageable is by attacking variety, said Mr. Kothari. This is because it is hard to imagine that future finance data sets will have less volume or less velocity than they do today. Structuring disclosures so that they are machine-readable facilitates easier access and faster analyses that can improve decision-making and reduce the ability of filers to hide fraud. Structured information can also assist in automating regulatory filings and business information processing. By tagging the numeric and narrative-based disclosure elements of financial statements and risk-return summaries in XBRL, the disclosure items are standardized and can be immediately processed by software for analyses. This standardization allows for aggregation, comparison, and large-scale statistical analyses that are less costly and more timely for data users than if the information were reported in an unstructured format. Structured data will likely drive future research in corporate finance and macroeconomics.

Another common big-data problem is accurately and timely connecting disparate big data sets for analyses. This problem is exacerbated by the broad range of identifiers used by federal agencies. A recent report identified 36 federal agencies using up to 50 distinct, incompatible entity identification systems. These differences raise costs and burdens for both federal agencies and their regulated entities. The Global LEI is a unique 20-character alpha-numeric code that offers a single international connector for disparate big data sets while reducing the regulatory burden associated with each agency's unique identification system. The LEI includes "level 1" data that serve as corporate business cards and "level 2" data that show the relationships between different entities. Recently, SEC released rules that mandate the use of LEI when associated with security-based swap transactions. The LEI is now a component of mandatory swaps transaction reporting in the U.S., Europe, and Canada. Mr. Kothari believes that the full benefits of LEI have yet to be realized. As some companies may have hundreds or thousands of subsidiaries or affiliates operating worldwide, more benefits lie ahead as the LEI becomes more widely and comprehensively used. The LEI allows more transparency regarding hierarchies and relationship mapping. This will support better analyses of risks as they aggregate and potentially become systemic tween different entities. 

Related Link: Speech

Keywords: Americas, US, Banking, Securities, Big Data, Policy Challenges, Cyber Risk, Tagging, Reporting, Swaps, LEI, SEC