Data-driven regulation and compliance is the key to a successful future financial services industry.
With regulators being the ‘public champion’ for new data technologies, this paper explores the automation of regulation and compliance and areas that offer significant potential to transform services and support the work of regulators.
The data science technologies of artificial intelligence (AI), Internet of Things (IoT), Big Data and behavioural/predictive analytics, and blockchain are poised to revolutionize regulation and compliance and create a new generation of RegTech start-ups. Examples of current RegTech systems include chatbots and intelligent assistants for public engagement, robo-advisors to support regulators, real-time management of the compliance ecosystem using IoT and blockchain, automated compliance/regulation tools, compliance records securely stored in blockchain distributed ledgers, online regulatory and dispute resolution systems, and in future regulations encoded as understandable and executable computer programs.
Automation is all the rage but why is this happening and what are the benefits? In short, money – cost savings and greater efficiencies are both imperatives and key drivers. Indeed, the UK Chancellor Philip Hammond put it well in a speech to the second International Fintech Conference in London on March 2019. His speech covered many areas of financial innovation, but perhaps the most significant aspect from a compliance perspective was his announcement that the FCA and Bank of England are moving towards automating regulatory compliance.
The intended benefits of this automation will likely be reduced costs for financial services firms as well as the removal of a key barrier for FinTechs as they enter financial services markets. Indeed, the proposed automation is part of the government’s new Fintech Sector Strategy, which seeks to retain the UK’s position as ‘the global capital of FinTech’ well beyond Brexit.
The UK has reason to be proud of its innovation. After all, FinTech contributes nearly $7 billion to the UK economy each year and London is home to 17 of the top 50 international FinTech firms. Indeed, in 2017, investment in UK FinTech more than doubled. For a country that launched the industrial revolution, evidently, innovation and entrepreneurship are far from over. Let us now translate some of these concepts into normal language by exploring their definitions.
Data Science and Technologies
In transforming regulation, the core technologies are:
• data facilities – online facilities of regulatory data collected by national government agencies, and often open-source for public access and analysis
• internet of things (IoT) – is the inter-networking of ‘smart’ physical devices, vehicles, buildings, etc. that enable these objects to collect and exchange data
• chatbots – systems for interacting with regulated companies, registrants and the general public using natural language and speech
• big data – is the process of examining very large data sets to uncover hidden patterns, unknown correlations etc.; data sets that are so complex that traditional data processing application software is inadequate to deal with them
• artificial intelligence (AI) – systems able to perform tasks normally requiring human intelligence
• behavioural/predictive analytics – the analysis of large and varied data sets to uncover hidden patterns, unknown correlations, customer preferences etc. to help make informed decisions
• blockchain technologies – the technology underpinning digital currency, that secures, validates and processes transactional data.
Regulators collect huge volumes of data (increasingly open sourced) and thus have major opportunities for so-called Big Data (analytics). In general, Big Data provides the opportunity of examining large and varied data sets to uncover hidden patterns, unknown correlations, customer preferences etc. Big Data encompasses a mix of structured, semi-structured and unstructured data gathered formally through interactions with citizens, social media content, text from citizens’ emails and survey responses, phone call data and records, data captured by sensors connected to the internet of things and so on. The notion of ‘Big Data’ is both increasing in volume, variety of data being generated by organizations and the velocity at which that data is being created and updated; referred to as the 3Vs of Big data.
Artificial Intelligence Technologies
AI technologies power intelligent personal assistants, such as Apple Siri, Amazon Alexa, and ‘Robo’ advisors, and autonomous vehicles. AI provides computers with the ability to make decisions and learn without explicit programming. There are three main branches:
• machine learning – is a type of AI program with the ability to learn without explicit programming, and can change when exposed to new data
• natural language understanding – the application of computational techniques to the analysis and synthesis of natural language and speech
• sentiment analysis – the process of computationally identifying and categorizing opinions expressed in a piece of text.
Behavioural and Predictive Analytics
Closely related to Big Data is behavioural and predictive analytics that focuses on providing insight into the actions of people. Behavioural analytics centres on understanding how consumers act and why, enabling predictions about how they are likely to act in the future, Predictive analytics is the practice of extracting information from historical and real-time data sets to determine patterns and predict future outcomes and trends. Predictive analytics ‘forecasts’ what might happen in the future with an acceptable level of reliability, and includes what-if scenarios and risk assessment.
Perhaps the most popular and much-coined term in FinTech, the core blockchain technologies are:
• distributed Ledger (DLT) – a decentralized database where transactions are kept in a shared, replicated, synchronized, distributed bookkeeping record, which is secured by cryptographic sealing. The key attributes are resilience, integrity, transparency, and unchangeable or mostly ‘immutable’
• smart contracts – are (possibly) computer programs that codify transactions and contracts which in turn ‘legally’ manage the records in a distributed ledger.
Automating Regulation and Compliance
A core focus of regulators of late has been the challenge of Digital Regulatory Reporting (DRR). Indeed, the Financial Conduct Authority (FCA) issued its Feedback Statement on its Call for Input FS18/2 in October 2018. This call for input outlines a ‘proof of concept’ developed at its TechSprint initiative in late 2017, seeking to make it easier for firms to meet their regulatory reporting requirements and improve the quality of the data that they provide. The feedback analysis weighs up the pros and cons of each of the following concepts, namely: disambiguation of reporting requirements; common data approach; mapping requirements to firms’ internal systems; a mechanism for firms to submit data to regulators; utilising standards to assist the implementation of DRR; a common data model; application programming interfaces; DLT networks; disambiguation of regulatory text; and of course utilising standards to assist the implementation of DRR.
In terms of the potential benefits of DRR, these can be summarised as follows:
• a reduced need for firms to interpret rules, making the information submitted to regulators more accurate and consistent
• increase in efficiency via reductions both in time and costs taken in complying with regulatory reporting requirements
• increase the attractiveness of the test jurisdiction’s regulatory framework for firms operating or considering operating in a certain jurisdiction
• increase the consistency of the information that regulators receive by reducing potential ambiguity within reporting requirements
• potential to implement future reporting requirements more quickly, and to improve the quality of data that are received were also commonly noted benefits
• provision of higher quality data is seen as a potential benefit for both regulator and regulated
• potential improvement in information sharing between firms – specifically internal risk and management purposes.
Regulation and Legal Status of Algorithms
Legal redress for algorithm failure seems straightforward. If something goes wrong with an algorithm, just sue the humans who deployed the algorithm. But it may not be that simple: for example, if an autonomous vehicle causes death does the lawsuit pursue the dealership, the manufacturer, the third-party who developed the algorithm, the driver, or the other person’s illegal behaviour? This stimulates the debate of whether or not algorithms should be given a legal personality in the same way as a company.
As we know, a ‘legal person’ refers to a non–human entity that has legal standing in the eyes of the law. A graphic example of a company having a legal personality is the offence of corporate manslaughter, which is a criminal offense in law being an act of homicide committed by a company or organisation. Another important principle of law is that of ‘agency’, where a relationship is created where a principal gives legal authority to an agent to act on the principal’s behalf when dealing with a third party. An agency relationship is a fiduciary relationship. It is a complex area of law with concepts such as apparent authority, where a reasonable third party would understand that the agent had authority to act.
As the combination of software and hardware is producing intelligent algorithms that learn from their environment and may become unpredictable, it is conceivable that, with the growth of multi-algorithm systems, decisions will be made by algorithms that have far-reaching consequences for humans. It is this potential of unpredictability that supports the argument that algorithms should have a separate legal identity so that due process can occur in cases where unfairness occurs. The alternative to this approach would be to adopt a regime of strict liability for those who design or place dangerous algorithms on the market, to deter behaviours that appear or turn out to have been reckless. Is this a case of bolting the door after the horse has escaped?
The prolific short story writer, Saki (H.H. Munro), used the term: “Design in haste, repent at leisure”. It is fair to say that a number of key regulators, such as the FCA, ASIC, Monetary Authority of Singapore (MAS), and the BVI’s Financial Services Commission are backing the future and facilitating innovation fast but their high level of consultation with key industry players means that repentance is unlikely – and why, with such cutting edge initiatives, should it be? We all know about tempus fugit and time flying but some of these leading regulators may just be more about carpe diem or seize the day. Long may it continue. If robots do their jobs, then perhaps some complacent regulatory noses deserve to be put out of joint and it may be appropriate to ‘cock a snook’ at the also-rans. Time will tell.
Professor Phillip Treleaven is the Chair of computer science at University College London and Simon Gray is Head of business development and marketing at BVI Finance
All credits for this story to the source below: