Supervision in a digital world: how modern technology is driving change
Speech by Pentti Hakkarainen, Member of the Supervisory Board of the ECB, at the EBI Policy Conference “Banking in Europe: a political, a monetary and a supervisory perspective”
Frankfurt am Main, 14 November 2019
I am often asked: how can you, as a supervisor, stay on top of technological changes in the banking sector?
This is indeed a relevant question, as it can be challenging for supervisors to understand and analyse new cutting-edge practices when these are being pioneered only in the private sector.
For this reason, our approach is not only to understand the implications of digitalisation, but also to begin using this technology ourselves in-house. I would like to take some time to explain how we in ECB Banking Supervision are using technology to our own advantage.
The ECB’s internal application of new technology
The incentives to adopt the most modern technology are equally strong in the private sector and the public sector. Adoption happens because there are opportunities to make work processes more efficient. In turn, efficiency gains allow us to optimise our use of resources – and thereby make us more effective in achieving our objectives.
As a modern supervisor, the ECB is investing in new technologies.
Our objective with this technological investment is not to replace supervisors with machines. On the contrary, the goal is to free up the time otherwise spent by our supervisors on routine and repetitive tasks – allowing them to focus on more value-adding activities that engage their cognitive abilities. As shown by the Financial Stability Institute’s latest survey on Supervisory Technology (SupTech), authorities think that the best results are achieved when technology is used to inform and support the expert judgement of supervisors. The idea is for humans to work together with machines; naturally with the humans remaining the master.
The benefits are not only for us as supervisors. There will also be benefits for supervised banks, as intelligent use of modern technology will increasingly allow us to streamline supervisory reporting and, consequently, to reduce administrative burdens.
Let me now move on to outline a few examples where we are deploying modern technology in the ECB’s supervisory work.
Natural language processing can be applied in our work to help us analyse unstructured data. Supervisors receive significant amounts of information in the form of reports and texts, such as annual reports, management information, and capital and liquidity assessments. We can use natural language processing to structure this information and to assess it.
As an example, we are making swift progress with the development of a tool that automatically machine reads the questionnaires submitted to us as part of fit and proper applications. The automated assessment will immediately identify any “red flags” or concerns about an applicant’s credentials that would require further scrutiny. It will, for example, flag issues regarding insufficient years of banking experience, or the existence of an ongoing criminal prosecution against the applicant. This initial automatic assessment will help to reduce manual and repetitive work, thereby allowing staff to invest their time in types of work that benefit from their expert judgement.
Structuring the data will also facilitate advanced analysis that can deepen our understanding of the issues that are relevant to us. For example, we want to deepen our understanding of how differences in board members’ time commitments or differences in the gender balance of boards can affect the governance and performance of banks.
Advanced analytical models are being deployed inside the ECB so that we can extract better insights from the data we hold. The amount of data available to us is vast, and fully comprehending the content may be beyond the capabilities of traditional analytical techniques.
To take one example, network analysis is being exploited as a means to understand private equity firms’ ownership stakes in supervised entities. Private equity groups are increasingly investing in European banks, with certain groups accumulating multiple holdings across a range of banks. To understand the implications of this, our analysts need to understand who owns the private equity groups concerned, and what cross-party participations they hold. This can be a challenging task, as the structural set-up of these funds tends to be complex.
To unravel this complexity, we use network analytics. This supports our supervisors’ analysis by providing an intuitive visual representation of ownership networks. The work has so far allowed us to analyse in more depth the bank holdings of over 300 private equity fund entities, capturing more than 150,000 inter-connecting investments. Assessing this complex picture in an accessible, visual way allows repetitive manual work to be avoided and helps our supervisors exercise their expert judgement.
Technology will also facilitate improvements in the way we communicate with the banks we supervise. As a first step, we are developing an online portal that will serve as a single gateway for our secure communications with supervised institutions. This tool will provide a user-friendly digital platform through which supervised banks will submit their applications for authorisation directly to the ECB. These applications will then automatically enter an integrated system, thereby supporting the collaborative work carried out with the national competent authorities (NCAs) on authorisation requests.
The portal will serve to reduce administrative burdens and to make our processes more transparent for supervised entities. Banks will be able to track the progress of their requests through the portal and keep up to date with relevant developments. The portal will also reduce the operational risks associated with the handling of sensitive information. Automated processes within this tool will bake in various checks and balances that prevent mistakes being made inadvertently. Also, the submitted data will be structured in a standardised way in our system. This will make horizontal analysis easier – thereby allowing us to carry out more rigorous benchmarking of the banks in our jurisdiction.
In time, we envisage that the portal will be extended to cover other supervisory processes and to bring Artificial Intelligence into the interaction and information exchange between supervisors and supervised banks. The platform will also help us in preparing conclusions and supervisory decisions.
International collaboration with supervisory partners
To enhance the speed at which supervisors can adopt value-adding technological innovation, close collaboration between international stakeholders is needed. As we are all experimenting with new technology, sharing the lessons we have learned helps others to avoid the pitfalls and maximise the benefits.
Conveniently, the structure of European banking supervision ensures seamless collaboration between our member institutions. This collaborative working arrangement between countries in the banking union is an essential design feature of European banking supervision, and it means that lessons learned are naturally passed on to all participants as a matter of course.
At a global level, we also actively exchange experiences and knowledge. In recent months we have explored the lessons learned by colleagues from the Monetary Authority of Singapore, the Prudential Regulation Authority in the United Kingdom and the Federal Reserve Bank of San Francisco. They have each pursued a number of technological pilot projects, and the findings from their efforts to demonstrate the viability of new technological ways of working have been very valuable to us.
Ongoing supervision of technological change in the private sector
In the final section of my remarks, let me return to the initial question: how do we in ECB Banking Supervision analyse and assess the technological change that is occurring in the private sector?
Part of this challenge is to have the right expertise within our organisation. As the previous section has demonstrated, many of our staff members have already developed strong technical competencies. This has been cultivated by our managers, who rightly encourage team members to take up training opportunities in this area.
I believe the ECB’s evolutionary approach to growing our technological capacity through supporting training and on-the-job technological experimentation is the best way to stay at the cutting edge. This is certainly a more sustainable approach than relying heavily on purchasing external expertise, and it also represents better value for money.
Now let me touch upon some technologically relevant elements of our supervisory activity.
One key aspect of technological change is the entrance of highly technologically sophisticated new banks into the market. As the supervisor for the euro area, it is our job to authorise the licence applications of all new credit institutions.
In this task, our guiding principle is technological neutrality. This means that our duty is to provide a level playing field for banks across the market, irrespective of the business model or technological strategy that they are pursuing. What counts for us is that prudential soundness is maintained, and this is what we seek to ensure when considering any licence application.
This does not mean that our analysis of each and every licence application follows an entirely identical process. On the contrary, we take a proportionate approach – one that is honed to the size, business model, and technological strategy of the institution involved.
In this context, in 2018 ECB Banking Supervision published a guide for the assessment of licence applications from technologically advanced “fintech” market entrants. The guide provides practical guidance to support NCAs in their assessment of and response to specific fintech-related risks. In turn, this helps to ensure a harmonised approach across European banking supervision when considering licence applications from fintech firms.
Having a document on this topic that is publicly available also provides transparency to the industry. This helps new banks that use fintech to better understand the most relevant procedures and the criteria that the ECB applies when assessing licence applications.
Whilst these technologically advanced entrants are an important element of the European banking market, their overall footprint in the sector remains relatively modest. Internal analysis of digitally oriented less significant institutions shows that they had total assets of around €30 billion at the end of 2018.
By contrast, significant institutions across the banking union accounted for €23 trillion of assets at the end of the second quarter of this year. This statistic reminds us how crucial it is for supervisors to keep abreast of the ongoing digitalisation processes in larger, incumbent European banks.
I can assure you that this is indeed what we are doing.
A multidisciplinary team of experts from all ECB Banking Supervision business areas is developing an approach to the supervision of banks that use fintech. This draws on the same pool of expertise that we utilise for implementing our internal technological initiatives, as outlined earlier in these remarks.
Fintech is a moving target and the ways in which banks are applying various technologies are still evolving. Developing the right approach to supervising fintech activity therefore requires supervisors to be flexible. It would not be helpful for the supervisor to jump to conclusions too quickly and to set rigid expectations that may not reflect best practice.
Instead, our aim in the first instance is to raise awareness among our supervisors of the relevant technological issues, and to develop a common understanding of fintech-related risks across European banking supervision. To achieve this we have opened a communication channel with banks to learn how they are using innovative technologies and solutions and what the implications are for their business models.
Regular contact with the industry is essential to keep abreast of the latest developments in innovation. In May we hosted an industry dialogue with a number of banks, by invitation. This event provided an opportunity to exchange views on issues arising from the use of fintech by banks. In particular, discussions focused on: credit scoring using artificial intelligence and big data, robo‑advice, and cloud services.
These exchanges have fed into the ECB’s supervisory approach for assessing how banks change their business models following the application of fintech and for mitigating related prudential risks.
In addition, the industry dialogue is informing our analysis of the prudential risks that are relevant in each new technological area. By deepening this analysis over time, we can become more specific in detailing what our supervisory expectations will be in our ongoing supervision.
These expectations boil down to common sense. Supervisors must ensure that banks have robust risk management frameworks in place vis-à-vis fintech. In this vein, we must ensure that banks’ business models are sustainable – also in the light of the technological strategy adopted.
It is clear that digitalisation also implies some new risks for the industry. All those involved share a common interest in controlling these risks, as it is in nobody’s interests for cyber-attacks on the industry to succeed. The public sector and private sector have therefore adopted a collaborative approach that seeks to ensure resilience – and thereby protect trust in the system.
Similarly, supervisors are working with the industry to ensure banks’ outsourcing of services to technological suppliers remains consistent with verifiably high prudential standards. This includes keeping an eye on the risk of excessive concentration in the supply of certain outsourced services across the sector.
To summarise, technology is at the top of the ECB’s agenda – it is at the heart of everything we do. Our work in this area is not something that we are just now starting, only to stop in a year or two’s time. It is a permanent, ongoing process to improve our efficiency and effectiveness.