We’re only in January and the FCA and Bank of England have already pulled their first regulatory rabbit out of the hat: a complete overhaul of their data strategy from 2013.
In a recently distributed press release, the FCA and BoE have acknowledged that in the seven years (!) since their initial data strategy was created that technology and advanced analytics techniques have developed at a rapid pace.
With the regulators having the unenviable task of trying to regulate an industry at a speed quicker than it's evolving, they have announced a significant plan of attack.
What are the regulators doing?
The FCA and BoE are looking to deepen their understanding of market functions, especially when it comes to the use of data collection and reporting practices (a part of the industry that has undergone considerable change in the past few years). In particular, the FCA is clear that over the next five years it wants to become a 'highly data-driven regulator'.
Inspired by a ‘Future of Finance’ report by Huw Van Steenis, the FCA in particular is planning to invest more in skills and resources in how it works with this side of the industry. With several working groups being formed across the regulatory bodies, the FCA and BoE have also issued a 56-page discussion paper entitled ‘Transforming Data Collection from the UK Financial Sector’. The paper is open to feedback from industry participants until April 2020.
What are the regulators looking at?
Essentially, the FCA and BoE want to reform how it goes about collecting and monitoring data from regulated firms over the next decade. The data the regulators gain from the industry play a crucial role in supporting supervisory judgements and in order to support this reform, a greater understanding is required for how the industry generates, collects and reports data.
Over 56-pages, the joint FCA/BoE discussion paper covers a number of topics which is fitting as a huge amount of data is being generated every day in all parts of the industry. In particular, the FCA is keen to understand how firms are already collecting data and wants to address ‘heterogeneity’ of data gained from firms with each organisation holding and describing equivalent data differently. This in turn makes it hard for the BoE to write a set of reporting instructions that are clear and unambiguous for all firms.
In addition, the paper is targeting heterogeneity of the BoE’s data needs as the regulator re-evaluates where data collected can be used for additional reports and overall making the data collection process more efficient and fluid. And, linked to heterogeneity of data submitted by firms, potential duplication of processes is also in the crosshairs. The regulators want to learn more about this and explore the potential for ultimately centralising more of these processes.
What could we expect to change?
The scope of the FCA/BoE’s work is wide and they're embarking on a considerable process, with the 56-page discussion paper covering a myriad of themes and potential solutions. However, there is a running theme which suggests the regulators has a potential favourite idea in mind in how to tackle these issues.
With the issue of heterogenous data clearly identified, many of the challenges the regulators face come down to how they can continue to efficiently and accurately collect data from such a wide array of sources. At several points in the paper, the FCA/BoE identify the potential to establish ‘common data inputs’ and ‘industry data standards’ to help move towards a common language for the industry.
On top of this, the regulators repeatedly raise the issue of adopting a ‘pull’ model where the BoE would be able to query data held by firms and generate reports on demand. In a format already seen in Rwanda and Austria, the regulator would essentially be able to ‘pull’ data it requires, increasing the speed of its data collection and potentially lightening the report generation burden on regulated firms. In Austria, this model works with the use of a ‘central service provider’ – a separate entity that collects and consolidates data from regulated firms for the benefit of the regulatory bodies.
However, in the discussion paper the FCA/BoE are aware that such a system comes with potential security risks and a ‘pull’ model would require a continuously working and reliable API. In order to deliver this, a certain level of investment and ongoing resource would be required as well as changes to governance. Overall, such a system drives at the notion of requests and processing of data becoming ‘machine executable’ which is something raised in the ‘Future of Finance’ paper. It is acknowledged in the discussion paper that, especially in times of crisis, regulators need to have immediate access to data to help form their views and make quick required decisions.
And at the same time, the BoE is exploring the use of natural language processing (NLP) of less processed and unstructured data from firms, called management information (MI). Such data could be analysed in powerful ways, such as comparing the extent to which boards of different firms are concerned about the same issues, and identifying inconsistencies between firms’ regulatory reporting, MI and published accounts. Promising, but in its current state this will probably only supplement - rather than substitute - structured data collections.
Another particular area the regulators are evaluating is how they draft instructions and guidance in an unambiguous format, translating this from natural language. Legal advice is being sought as they explore moving towards precise instructions for how data is handled and submitted and it is acknowledged in the discussion paper that writing such instructions as code poses significant technical and legal challenges.
Other changes to reporting instructions, not dependent on common data inputs, could involve annotation to make instructions easier to navigate, use of more standardised language to improve precision, or deeper collaboration with firms when developing instructions.
How could this impact you?
In the immediate future, it's realistic to say the above will have little impact on regulated firms. The regulators aren't known for being fast acting and with the discussion paper remaining open for several months, they will undoubtedly want to take their time with an area as fluid and complex as data collection.
However, in the medium to long term it's clear the FCA and BoE are serious about how they go improving their activities in this field and even though the cogs may be slow to turn, change will inevitably come.
The outcomes of the discussion paper are in the balance but firms could expect to receive new guidelines on how data is generated and consolidated for reports. Uniformity of data will make things a lot easier (and cheaper) for the regulators to collect and in time a system similar to Austria, where a ‘central service provider’ is used, could be adopted.
And as the FCA continues to invest in resources, and establishing internal working groups in this area, its understanding of data and how it is generated will only improve. As a result we could be set for a more educated regulator as it becomes a more engaged one. Which, if at least for the benefit of end consumers , can only be a good thing.