What is digital regulatory reporting and why should you care?

The UK regulators are testing whether compliance with their rules can be automated. In the future, reports that regulated firms must send to the regulators could be generated and submitted without needing human involvement. However, there are significant hurdles to overcome before automated regulation becomes reality.

What is regulatory reporting?

It is fair to say that reporting is not the most glamorous area of financial services regulation. But, thanks to the work of the FCA and Bank of England, it has recently received a lot more attention as a pioneering example of regtech. It even made it to top of the list of actions in the UK Government’s Fintech Strategy.

Regulators require firms to submit huge volumes of data to them (for example about their business, products and transactions). The problem is that each firm interprets those rules in different ways. As a result, the data that the regulators receive is not necessarily consistent across the industry.

To tackle this, the FCA and Bank of England are running a project that they are calling “digital regulatory reporting” (previously known by the less catchy title “model driven machine readable regulatory reporting”). This project is exploring whether technology can make submitting data to the regulators less reliant on human interpretation of rules.

What are the regulators doing?

The regulators ran a techsprint last year on automating regulatory reports as a proof of concept. A techsprint is the FCA’s version of a hackathon i.e. bringing together industry participants to collaborate on finding a technological solution for a given problem in financial services.

The techsprint concluded that it is possible to write rules in a language that could be read and executed by a machine once it was mapped to data held by firms. It also demonstrated that a change to the machine-readable rules would update the data being reported in real time.

As a next step, the regulators are now running two six-month pilots with industry participants to test the scalability of the project. The FCA has also published a Call for Input to get feedback.

What are the legal challenges?

To make rules machine-executable, first you must make them machine-readable. And, because machines read in a different way from humans, this means using a less ambiguous form of language. Machines deal in black and white rather than the shades of grey sometimes preferred by lawyers and law-makers.

However, whoever is doing that rewriting will need to make sure that their version of the rule still captures all the potential ambiguity of the original. If not, it is possible there might be a gap between the requirement and the interpretation of that requirement.

Some of the questions arising from this are:

  • Who is responsible for creating the machine-readable interpretation of the rules?
  • Would they be responsible for getting that interpretation right?
  • How confident would regulated firms be that applying that interpretation?
  • How will that interpretation be future-proofed or maintained?
  • What is the regulators’ role in this going to be?

We await to see how these legal challenges and other issues are addressed in the Call for Input and outcome of the pilots. In the meantime, reporting continues to be a somewhat manual labour.

See also our animation on the digital regulatory reporting project.