A recent FSB report offered case studies of how regtech and suptech are being deployed by regulators, including several examples from MAS.
Recently the FSB (Financial Stability Board) published a new report on the use of SupTech and RegTech, outlining the opportunities, benefits, risks and challenges they present to authorities and regulated institutions alike.
The report offered 28 case studies providing practical examples on how technology is being deployed by regulators, including several examples from MAS (Monetary Authority of Singapore).
Below is a summary of MAS’ use of technology to enhance its regulatory and supervisory functions, as outlined in the FSB report.
Case study: Network analysis for STRs
To enhance the effectiveness of its AML/CTF supervision, MAS has developed an STR (Suspicious Transaction Report) network analytics tool which it uses to better analyse FIs and identify concerning clusters of individuals/entities that exhibit suspicious behaviour.
The insights and emerging risks uncovered from the network analyses help MAS target higher risk areas and FIs for closer supervisory scrutiny. The analysis is also shared with the broader financial sector to aid the national effort to combat financial crime.
The data inputs for the network analysis in the initial phase comprise mainly of information from the structured data fields in the STRs.
However, the dataset is being enhanced to process more transaction data and company profile information, as well as to include NLP (natural language processing) models which will extract information from unstructured textual data within STRs (e.g. narratives).
Case study: Predictive modelling to identify misconduct risk
MAS has developed a simple multi-factor logistic regression model that can predict the risk of misconduct for financial advisers over a period of two years. The model uses different factors, such as working experience and misconduct history, to predict future misconduct.
Using the model, MAS is able to identify representatives and transaction samples for scrutiny during onsite inspections.
Case study: Text analysis of audited financial statements
MAS has deployed a tool that uses text analysis techniques, combined with quantitative analytics of financial metrics, to analyse audited financial statements submitted by FIs to identify areas of concern.
MAS is able to visualise the results of the analysis on a single automated dashboard, giving supervisors a bird’s eye view of the reports. The tool has improved the efficiency of MAS’ review of audited financial statements and helped supervisors identify and focus on red flags more quickly and systemically.
The tool has been deployed as a pilot in one department so far, but MAS is looking to widen its use to more supervisory departments.
Case study: Data analytics for inspections
MAS has developed a tool that automates the process of reviewing firms’ trade data and identifying potential red flags, which has traditionally been a manual process that relied on sampling.
The tool leverages algorithms and statistics to analyse entire datasets – such as when reviewing trade allocations and prices – allowing supervisors to identify specific trades of greater concern and focus on statistical outliers.
Market-basket analysis and other techniques are also used to identify accounts that frequently trade together.
Case study: Monitoring and enforcement of safe distancing measures
MAS used data analytics to monitor regulated FIs’ implementation of safe distancing measures amid the Covid-19 pandemic and inform inspection and enforcement actions. Data on bank branch locations, customer footfall, wait time and peak hours are collected and visualised on a monitoring dashboard.
The results are used to prompt intervention actions and prioritise inspections to enforce compliance with safe distancing rules.
Other Covid-19 related initiatives
MAS deployed automation tools using NLP to gather international news and stay abreast of Covid-19 related developments.
NLP was also used to analyse consumer feedback on Covid-19 issues, and monitor vulnerabilities in the different customer and product segments.
MAS also collected weekly data from regulated institutions to track the take-up of credit relief measures as the Covid-19 pandemic unfolded. Data aggregation and transformation was automated and visualisation allowed MAS to identify pain points and issues for policy analysts to examine in detail.
MAS is also exploring the development of an integrated surveillance platform to collate and aggregate data from various sources, apply NLP/sentiment analysis, and enable risk identification using machine learning.
The regulator is also exploring the use of rule-based and AI/machine learning techniques to estimate credit gradings of loans based on factors such as financials, adverse news, account conduct and covenant breaches. This will enhance MAS’ surveillance and credit risk monitoring capabilities, and allow onsite inspections to be conducted more efficiently.
The FSB report, available here, also includes examples from regulators in the EU, UK, China and the US.