Certain challenges in securities finance can only be met with better data and newer data models. Market regulators now coping with investor demands for ESG-compliance will have to monitor the disclosures of regulated entities by combing through vast pools of stock loan and proxy voting data. Bank custodians and brokers, if tasked with validating the social propriety of their stock loans, will have to dive deep into customer profile data, deeper than either regulators or vendors can today access efficiently.
Distributed ledger technologies may offer solutions -- if that confidential data can be mobilized and still protected at the same time. We believe the use of data trusts can solve that confidentiality problem for both regulators and market participants. The time has come for beneficial owners, borrowers, and their service providers to bring their transaction details and unstructured policy-level records under one roof.
The key attributes of a data trust are ownership and control.
A data trust is an innovation listed by the Massachusetts Institute of Technology as one of the ten most significant breakthrough technologies of 2021. Its legal structure fuses contemporary notions of privacy controls on "Big Data" with traditional trust laws and fiduciary duties. [1]
Every data trust's central organizing principal is that the trustees are instructed to use their data assets for the owners’ exclusive benefit. To achieve that purpose, the trust owners define rules for data usage. With those policies in place, the governing body takes responsibility for enforcing access security. As in most institutional trusts, a board of trustees usually outsources the asset (data) safekeeping and delegates its management to trusted contractors.
In the near future, forensic and data analysts will use these data assets within trusts to deploy smart contracts to evaluate credit access and perform otherwise impossible tasks. Their algorithms and oracles will help to categorize loans, rate policy documents, and then post encrypted transaction records to shared ledgers.
Securities loan data is an inherently valuable asset.
Transaction feeds from service providers provide the raw data for commercial benchmarking vendors. Today, lenders have no choice but to give data away to those vendors - then buy it back through "peer reviews." Moreover, much of the data's value is left on the table. Why? Because most of the transaction and policy data remains siloed in proprietary systems that do not share well.
Data feeds can become powerful tools if the new distributed ledger technologies are employed in its usage. What if that siloed data can be anonymized, then securely combined with other funds' data? The resulting data collection could span the entire industry and support hitherto unimaginable risk management tools. Such a compilation would then also enable the owners to maximize their own data assets’ value without sacrificing their intellectual property rights or revealing their competitive strategies.[2]
Both the public and private sectors are embracing data trusts.
Many data trusts have been government sponsored, like Virginia's Commonwealth Data Trust. With over 2,000 operational data systems, Virginia's data trust controls everything from library data to crime, corrections, EMS, hospital, patient, and aviation records. Similarly, private data trusts such as that of the Mayo Clinic are designed to collect all "data from patient care, education, research, and administrative, transactional systems, [that is] organized to support information retrieval, business intelligence, and high-level decision making."[3]
Although a data trust for securities lenders and borrowers would be an original application of the concept, the European Mastercard data trust may provide a useful precedent. Truata was formed to anonymize customer transaction data for analysis and compliance with the EU's strict consumer privacy regulations. Truata's beneficiaries are competitors, just like securities finance market participants, so they rely on robust usage and encryption policies that make it difficult for owners to use the data as a weapon against one another.
Data owners design their own data trusts.
A data trust doesn't necessarily have to take the form of a trust in the legal sense. It can be a limited liability company, partnership, or corporation, whatever legal form the data owners decide works best for their purpose. The same is true for the trust's governance structure. The data owners have a great deal of flexibility in structuring the trust, so its governance and management reflect the consensus of the data owners and their purpose for combining their data. Despite this inherent flexibility, there are some basic things every data trust must have in place, regardless of the legal or contractual methods employed to achieve its purpose.
Why not use commercial data vendors' systems? They aren't equipped for compliance and surveillance work. As we've discussed previously, the securities finance databases of leading data service providers were designed as long as 20 years ago, mainly for performance benchmarking. The lending data these firms receive from intermediaries and funds are insufficiently granular to be used either for cross-border compliance monitoring or for proxy voting metrics. Their systems were not designed to hold the large amounts of raw and unstructured transaction data that will be necessary to solve the challenges for securities finance.
The “give to get” model has a new competitor: the scalable, encrypted data trust.
An industry-wide securities lending data trust would release beneficial owners from the cycle of supplying data to aggregator firms, who then process it and sell it back to them. By contributing to a data trust, the lenders and borrowers would continue to benefit from benchmarking services from vendors, but by asserting ownership of their data, they could also use their resources to prove compliance with market practices and regulations.
A voluntary industry-wide securities lending data trust would provide a scalable alternative to today's ad hoc and incomplete data sharing. Under common ownership, and employing common rules for encryption, data security, privacy, anonymization, and confidentiality, securities lending market participants can finally get the full value out of their data resources.
In Part II, we’ll discuss in more detail how securities lending industry participants can see real value by contributing to a data trust and how a data trust can foster best practices and ease regulatory burdens by automating the work of regulators.
Notes:
[1] A "data trust" is established when separate entities place data under the control of a board of trustees (or other governing body) with a fiduciary duty to manage and safeguard the data in the interest of the data owners.
[2] Data can also be a weapon that can be used by competitors and by litigious investors. A carefully crafted data trust can ensure that data is properly controlled and handled with the utmost care and discretion.
[3] Other examples of data trusts include: