Chapter 2: RegTech, SupTech and the future of compliance

Question 1

What caused the heightened regulatory expectations?

Answer 1

The aftermath of the 2008 financial crisis has led to much-needed
regulatory reforms, large fines for misconduct and mismanagement,
and heightened compliance and risk management expectations within
firms.

Question 2

What is RegTech?

Answer 2

RegTech is generally defined as the adoption of new technologies to facilitate more efficient and effective delivery or regulatory requirements.

Question 3

What does RegTech do?

Answer 3

reduction of the cost of compliance via automation, or leveraging technology to increase the effectiveness of compliance (for example, by accessing broader data sets or employing better data analytics).

Question 4

Who are the key players in the RegTech ecosystem? (4)

Answer 4

1. RegTech companies,
2. regulators,
3. financial institutions and
4. professional service providers, such as accounting,
   legal, compliance and tax experts and advisers.

Question 5

What are 3 types of RegTech application which the FCA has identified?

Answer 5

• those aimed at supporting regulatory compliance firms
• those aimed at improving regulatory oversight and modernising
regulation
• those aimed at re-engineering or reforming regulatory systems.

Question 6

What is SupTech?

Answer 6

RegTech’s counterpart, SupTech, is revolutionising the work of supervisory agencies.

Question 7

What does SupTech do?

Answer 7

Instead of periodically collecting aggregated data in reporting templates, ‘data pull’ approaches source data directly from the operating systems of regulated institutions at intervals ranging from 24 hours to 15 minutes.

Question 8

What are the advantages of SupTech? (4)

Answer 8

This allows for the real-time monitoring of transactions, minimising reporting errors, and removing the opportunity for financial misstatements – even allowing automatic
incorporation of changes in regulatory requirements into the technological reporting protocols.

Question 9

What are data-input approaches?

Answer 9

‘Data-input’ approaches, involve reporting institutions submitting data that are encoded into a human- and machine-readable format that use standardised electronic taxonomies, or ‘tags’ and sending it to a central database in a non-aggregated
form.

Question 10

Give an example of data consolidation and analysis based on SupTech solutions.

Answer 10

For example, the Bank of Italy is combining suspicious activity reports with natural language processing analysis of press reviews to detect money laundering.

Question 11

Give an example how the FCA uses trained algorithms.

Answer 11

the FCA has trained algorithms to model normal trading behaviour and automatically report signs of insider trading. These are revelations which otherwise might only have been brought to light by a human whistleblower, if at all.

Question 12

How can SupTech help governmental organisations?

Answer 12

SupTech can also be used to refer to the use of technology by governments
to supervise their own agencies, and to assist with the supervisory
activities of the public and independent watchdog organisations.

Question 13

What are challenges to RegTech adoption? (6)

Answer 13

1. Lengthy procurement process
2. Preference for large and established providers
3. Privacy
4. Bias decision making in ML
5. Black Box in using ML
6. Adoption of the Cloud

Question 14

Explain why privacy concerns has prevented RegTech from further adoption.

Answer 14

The first obstacle is the balance between transparency and privacy. A survey conducted by Ernst & Young revealed a ‘tension between opinions about what channels companies should monitor and the types of surveillance that their employees
consider a violation of privacy’. For example, around 65 per cent of respondents respectively felt that email and phone call monitoring was a violation of privacy.

Question 15

Explain bias in ML algorithms and give an example.

Answer 15

ML algorithms will learn from and perpetuate distortions in training data. Moreover, inherently algorithms are optimised to achieve particular goals, which can lead to biased decision making.

RegTech and SupTech are not immune. For example, fraud detection algorithms have been shown to be biased against certain ethnic minorities, immigrants and even against men.

Question 16

What is a counterargument for ML being capable of being bias?

Answer 16

the question arises: How much less biased than a human does an
algorithm have to be before we are willing to let it loose?

Question 17

Explain the Black box principle.

Answer 17

Another obstacle of ML is that it is not clear how or why particular classification has arrived from the inputs.

Question 18

What can be expected in the future with the financial services industry becoming increasingly digitalised and data-driven?

Answer 18

we can expect the regulatory structure to shift from a rule-based regulatory framework to a more data-driven and proactive approach to compliance and supervisory monitoring.

Question 19

Identify 6 RegTech use cases.

Answer 19

1. Identity management & control
2. Risk management
3. Regulatory reporting
4. Transaction monitoring
5. AML transaction monitoring
6. Trading in financial markets

Question 20

Based on the BIS report, in which areas of financial supervision is SupTech commonly used? (7)

Answer 20

• automated data reporting
• data validation focusing on the checks for receipt of data, checks for data completeness, checks for data correctness and plausibility and consistency checks
• virtual assistance (through the likes of chatbots) to answer consumer complaints and provide assistance to supervised entities automatically
• market surveillance focusing on suspicious trading, such as market manipulation and insider trading
• market surveillance
• misconduct analysis
• identification of micro- and macro-financial risks.