In this article we look at BearingPoint's predictions for the evolution of regulatory technology and the Financial Conduct Authority's latest evaluations of artificial intelligence or AI.
BearingPoint unveils its 'RegTech Factory'
BearingPoint RegTech, the well-known provider of supervisory and regulatory technology, is now marketing a piece of outsourcing software that produces regulatory reports; the firm hopes that it will lower the cost of compliance at institutions.
BearingPoint blames the spiralling cost of compliance at financial firms squarely on the fact that most of them still rely on manual processing. It exempts "challenger banks" from this accusation because they are not old enough to be operating antiquated IT.
By way of example, BearingPoint states that the Financial Conduct Authority has guessed that the annual cost of regulatory reporting at a typical large regulated mortgage firm is £450,000 in terms of continuing costs and a further £700,000 in terms of change-related costs (in addition to technological costs, which the regulator has excluded). The meaning of these terms is vague.
In response, BearingPoint’s ‘RegTech Factory’ provides firms with a standardised granular data model with which they can interact (or, to use the firm's terminology, "interface"). These firms, if they sign up to it, are therefore destined to share the same infrastructure, while moving to a more granular data model than before. This, according to BearingPoint, will shield them at least slightly from regulatory change. It describes the new system as "regulatory Reporting-as-a-Service."
Mark Shead of BearingPoint RegTech said: “The UK banking industry is competing in processes where no revenue is being generated. We believe that this situation will change. There is room for financial institutions in the UK to work together more, collaborating on the conceptualising of reporting tools, the resultant IT build and the deployment of new reporting systems.”
BearingPoint refers to this collaboration as "the economies of share" and it is worth us taking a cursory look at its vision for RegTech in the immediate future.
The three factors
The firm believes that three major trends will shape banking regulation in the next five to ten years:
- granularity, near-to-real-time reporting and triple-entry accounting;
- the standardisation of data models and data processing logic; and
- "co-opertition" (co-operation and competition between two firms in different areas simultaneously) in the financial sector in general.
Governments require banks to report data in templates at a set frequency, having sorted and aggregated the data according to various set criteria. BearingPoint sees a clear trend away from the collection of aggregated data and towards requests for granular data sets which firms must report promptly and "validate" in great detail, an example being transaction-based reporting for derivatives and AnaCredit. According to BearingPoint, the average regulator is going to become more and more interested in receiving granular information from two counterparties in such a way that both sets of data complement one another. This is known as triple-entry accounting. In the triple-entry accounting system, all accounting entries are cryptographically sealed by a third entry and this deters manipulation and financial fraud.
Regulators have to set clear standards for data models and the associated data processing logic if they want to receive granular data promptly. Despite many projects such as the EU's BIRD and IReF and a clear mandate from that body as expressed in Article 430(c) of its second Capital Requirements Regulation, progress here has been modest. BearingPoint writes: "Currently, data is queried several times for different reporting frameworks (e.g., liquidity, solvency, statistics). Instead, at least one output data format should be specified that defines the necessary information for a business data set with which the regulator can perform all its tasks within the existing regulatory frameworks. In many cases, it would even be useful to define an input data format based on which data can be uniformly processed."
Banks compete in both core and peripheral businesses. Competition between them in their core business ought to earn them the most revenue, but competition in other areas of business is often a waste of time. BearingPoint makes the obvious point - held by this publication but, unfortunately, not by any known bank - that regulation is one area in which banks ought not to view each other suspiciously. For this reason, BearingPoint wants to see a state of "co-opertition" - in other words, competition in core business and co-operation elsewhere, with banks sharing resources to reduce costs. Co-opertition (BearingPoint spells the word slightly differently) models are becoming more interesting in the financial sector as a result of digitisation, the expansion of Cloud infrastructures, the increase in computing and storage capacities and the never-ending development of platform business models. In the case of regulatory reporting, for example, banks could share the cost of IT infrastructure, IT deployment, IT maintenance, IT licences, regulatory analyses and IT design.
AI and the regulators
Meanwhile, representatives of the likes of Standard Chartered, HSBC and Credit Suisse attended the Bank of England's and Financial Conduct Authority's first Artificial Intelligence Public-Private Forum recently. Several attendees expressed the view that Covid-19 has accelerated the pace of automation and the adoption of AI in financial services. Covid-19 has also accelerated the shift towards online society and existence. Delegates thought that AI will probably become an increasingly integral part of this in the short-term and that firms therefore ought to consider AI as a socio-technical issue. If people adopt AI in the same way as they adopted the Internet, it was reasoned, there might be a gradual increase in the benefits associated with it, interspersed with periods of relative inertia.
The meeting disparaged the idea that financial firms ought to look at "digital-native companies" in other sectors to see how they design, test, implement and monitor AI applications. Devices that use AI in homes, workplaces and other aspects of daily life are all too often "going against principles of accountability, fairness, privacy and transparency." When a consumer plays a video or looks for a song on one of the many streaming platforms, AI is at work, recommending algorithms that offer the sort of content that suits him. The same goes for search engines, both those used in browsers and those found on e-commerce sites. Microwave ovens that use voice control instead of direct app control are examples of AI. Refrigerators and ovens can be operated by use of a mobile phone, also using AI, and Sous Vide cooking (heating food with water) uses it.
The FCA and The Alan Turing Institute are working on a year-long collaboration to promote "AI transparency," something that the FCA defines as "stakeholders having access to relevant information about a given AI system." This promises to be a major area of regulatory intervention in future. One benefit of well-regulated AI that the FCA sees is the ability of customers to understand and perhaps challenge the reasons why their banks make particular decisions, especially if they base them upon algorithmic assessments of such things as creditworthiness and anything else that might suffer from bias.