Algorithms have come in for some bad press lately, but they are only as good as the inputs and data they rely on. What outcomes do they deliver? What outcomes were intended? Do they exacerbate the biases inherent in society?
As Andrew Bailey hinted at in his speech on Stablecoin at the beginning of the month, the answer to new tech is not to pull up the drawbridge. Rather it is for the regulator to discriminate. To do that effectively, we need a fuller understanding of the tech and its benefits, weighed against inevitable risk.
It was to develop that understanding that I proposed Project Innovate, the FCA’s regulatory sandbox and latterly the global financial innovation network of over 60 regulators and other organisations. Our aim is to better understand novel business models and tech usage, both to encourage new entrants and to ensure we have the knowledge base to rule against innovations we doubt are in the markets’ or consumers’ interests.
Faced with greater innovation at bigger scale, in my view the sandbox should be expanded in terms of scale and ambition to handle occasional very large tests, if need be backed with more flexible authorisations powers. Greater use could also be made of calling for specific cohorts to test answers to specific issues, which the FCA has already tried successfully. What we have also found is large datasets are often a barrier to new entrants being able to design and build products – the partnership with the City of London on a digital sandbox presents an answer to that problem.
FCA reform
Innovate was one example of how we have changed as a regulator and I was incredibly proud to have sponsored our work and built the team from an initial 2 to over a 100 colleagues focussed on fintech and our own use of innovative technology. The team has already helped 1,700 firms and our data science units won its spurs during the last six months as the need to develop urgent and rapidly updated information has grown.
But there are more ways the FCA is changing and needs to change to be ready for the future.
The first is how we use data and analyse the intelligence we receive.
You cannot hope to effectively regulate 60,000 firms, most of them small, as we do, unless we make progress on this front.
Last year, we put in place a new data strategy to make sure we brought in the analytical skills we need, to raise the minimum standard of data understanding among all our staff, to ensure we have the kit to take full advantage of the skills we have and the huge amount of data on which we sit.
Embedding this is well under way.
Continue reading article…
