The 2008 financial crisis and the 2023 US banking turmoil exposed the limits of existing supervisory frameworks in capturing systemic liquidity risk, particularly as digital banking and non-bank financial institutions reshape funding dynamics. This blog outlines a comprehensive SupTech-enabled framework for liquidity supervision, covering automated reporting, stress testing, network analytics, and the practical challenges of adoption for emerging market authorities.
Author: Pamela Kahwa, Financial Sector Risk and Policy Analyst, Bank of Uganda
This article represents the author’s personal perspectives based on research conducted during the Innovation Leaders Residency initiative.
_________________________
Introduction
For financial authorities, establishing a robust framework to assess liquidity risk in the financial sector is essential to ensure that supervised entities effectively manage their liquidity to meet obligations and withstand sudden funding shocks. Liquidity crises, whether idiosyncratic or systemic, can lead to the failure of financial institutions and trigger instability in financial markets, as demonstrated by the 2007 Great Financial Crisis (GFC) (International Monetary Fund, 2010). The GFC exposed vulnerabilities in the financial system that had been overlooked by supervisors, prompting a wave of regulatory reforms. These reforms introduced comprehensive laws, principles, and guidelines to enforce prudent liquidity risk management practices in banks and bolster the financial system’s resilience to liquidity shocks (Bank for International Settlements, 2018, Basel Committee on Banking Supervision, 2013).
While these reforms have significantly enhanced the global banking system’s ability to withstand severe crises, the United States (U.S.) banking crisis of 2023 underscored the ongoing importance of strong and effective liquidity risk supervision (Basel Committee on Banking Supervision, 2023). The crisis revealed two key developments that have changed the nature of liquidity risk in the banking system. First, the proliferation of digital banking technologies and the role of social media in amplifying negative news have contributed to the reduction in the stability of bank deposits by intensifying the risk of bank runs (Fernando Restoy, 2024). Second, rapid financial innovation has broadened the scope of market participants to non-financial institutions with unique business models, cash flows, and funding structures, thus changing the nature of banks’ liquidity exposures.
The lessons from the 2023 crisis demonstrate that the existing supervisory framework requires enhancements to enable more effective liquidity risk oversight, ensure timely intervention, and quickly adapt to financial system changes, consumer behaviour, and emerging risks (Group of Thirty, 2024). This article discusses the essential elements needed for a comprehensive supervisory framework for monitoring liquidity risk. The article also explores how financial authorities can leverage supervisory technologies (suptech) to enhance their capacity for liquidity risk monitoring.
Towards a comprehensive supervision framework for bank liquidity risk
Liquidity risk refers to the potential inability of a financial institution to meet its cash flow and collateral obligations — both expected and unexpected (Bank for International Settlements (2018), European Central Bank (2018)). It encompasses funding liquidity risk (difficulty in meeting funding needs), market liquidity risk (inability to quickly adjust positions due to insufficient market depth or disruptions), and systemic liquidity risk (simultaneous financial strain across multiple institutions, often linked to reliance on short-term wholesale funding). Effective mitigation requires both microprudential oversight and macroprudential interventions to address risks that extend beyond individual institutions. As such, effective supervision of liquidity risk requires a diverse set of tools, as no single metric can fully capture its complexity. The following are the proposed key considerations for a robust supervision framework for bank liquidity risk.
1. Comprehensive measurement and monitoring
Supervisors should account for the interplay between funding liquidity risk and market liquidity risk. Measuring liquidity from capital markets requires recognising their higher volatility compared to traditional retail deposits. Additionally, supervisors must consider various factors that threaten banks’ liquidity, including financial and operational risks, perceived or actual weaknesses, activities and strategies that may create significant liquidity strains, and events that could influence market and public confidence in banks’ soundness.
To comprehensively capture the different dimensions of liquidity risks, supervisors should evaluate balance sheet structures, projected cash flows, liquidity positions in major currencies, and off-balance sheet risks. Analysis of vulnerabilities should span normal and stressed conditions over various time horizons, identifying funding gaps and potential needs arising from projected outflows relative to available funding sources.
2. Integration of systemic dynamics
Existing measures like the Liquidity Coverage Ratio (LCR) and Net Stable Funding Ratio (NSFR) are insufficient for addressing systemic liquidity risk because they fail to consider institutional interconnectedness and contagion effects. Supervisors must analyse the factors driving market and systemic funding illiquidity, as well as the channels through which liquidity shocks are transmitted and amplified within the banking system (International Monetary Fund, 2010). This requires the ability to map and quantify interactions between banks and, increasingly, non-bank financial institutions (NBFIs) while monitoring structural changes in the financial network that affect systemic liquidity. The framework should facilitate system-wide analysis by examining the links between bank leverage, liquidity creation, and systemic risk (Acharya and Thakor, 2017).
3. Early warning systems and adaptive response
The appropriate early warning systems can help supervisors to identify potential funding needs or the emergence of increased liquidity risk or vulnerabilities in banks. Such early warning tools, qualitative or quantitative, should identify any negative trend and cause an assessment and appropriate response to prevent the materialisation of the risk or minimise its impact. The framework should also facilitate the adjustments of assumptions about market liquidity based on prevailing market conditions or institution-specific circumstances (Bank for International Settlements, 2018).
Leveraging suptech for monitoring and mitigation of systemic liquidity risks
Enhancing liquidity risk oversight
The complexity of an ideal supervision framework for bank liquidity risk necessitates an advanced implementation approach by financial authorities to address the challenges of this risk category. Supervisors currently use tools like the Basel III liquidity monitoring tools and stress testing models (Ong, 2014) to evaluate banks’ liquidity positions, assess their resilience to stress, and define enforceable regulatory limits. While effective for narrow scope application, these tools have limitations in enabling comprehensive, granular assessments. These gaps stem from insufficiently diverse data sets and limited capacity to develop robust liquidity models (Baudino et al., 2024). Recent developments underscore the need for supervisors to enhance their frameworks to capture the evolving dynamics of bank liquidity channels, improve the calibration of stress scenarios to reflect financial innovations and technology-driven structural changes, and incorporate interlinkages between banks and non-bank financial institutions (NBFIs). To meet these demands, financial authorities are turning to supervisory technology (suptech), which leverages innovative technological solutions to support financial system oversight (Beerman et al., 2021).
Suptech applications enable process automation, reduce regulatory data processing costs, and enhance analytical and decision-making capabilities (Broeders and Prenio, 2018). The State of SupTech Report 2024 by the Cambridge SupTech Lab (Simone di Castri et al., 2024) reveals that 65 percent of surveyed financial authorities utilised at least one suptech application in 2024, primarily for prudential supervision, consumer protection, market conduct oversight, and anti-money laundering/counter-terrorism financing (AML/CFT). These tools augment supervisory processes, benefiting regulators, supervised entities, and financial consumers alike. Suptech, therefore, can equip supervisors with dynamic tools to access diverse and granular data sets, extend the scope of liquidity risk monitoring across various financial institutions and instruments, and develop advanced analytical models. As such, suptech represents a pivotal advancement in strengthening supervisory capacity and addressing systemic liquidity risks effectively.
Proposed application areas
Supervisory authorities must put in place effective liquidity risk assessment frameworks that consider both short-term liquidity management and long-term funding strategies by supervised entities, incorporating stress testing and scenario analysis, and ensure strong governance and oversight to enable the detection and management of systemic liquidity challenges in an increasingly complex financial environment. This section outlines areas where suptech can be deployed for this purpose.
1. Automated reporting and data analytics
A robust supervision framework for liquidity risk should enable regulators to monitor diverse cash flows, institutional behaviour as per balance sheet exposure and market activity, inter and intra financial system interconnectedness, and measure the effects of exogenous factors on bank liquidity conditions. This requires high-frequency, granular analysis to properly detect anomalies and the build-up of weaknesses in individual banks and markets.
An example where the transition from aggregated data to granular data has facilitated continuous reporting and liquidity tracking is the Indonesia Financial Services Authority (OJK)’s OJK-Box (OBOX). The OBOX provides access to granular transactional data where previously only monthly aggregated data was received and additional data was requested from the financial institution. OJK’s supervisors can now easily detect irregular transactions and take immediate action to mitigate potential issues that may arise (International Financial Consumer Protection Organisation, 2020). In Nigeria, the Central Bank of Nigeria has developed the Agent Banking Management System (ABMS) to monitor mobile money transaction reports and performance across service types and digital channels for all agent locations. The supervision tool is designed to identify breaches and predict consumer behaviour using trend analysis (International Financial Consumer Protection Organisation, 2020).
Evidently, suptech solutions can help to automatically gather and process large volumes of granular, high-frequency data from regulated entities, ensuring timely insights and eliminating manual reporting delays. They can collect and integrate data from multiple sources, such as financial market infrastructures (FMIs), bank balance sheets, and transaction data from payment systems. Analytical tools can then be developed to use this data to track liquidity flows and identify stress points across various instruments, exposures, and maturities where disruptions in liquidity could indicate broader systemic issues. This enables central banks to monitor liquidity positions in real-time and assess indicators such as loan-to-deposit ratios, LCRs, and interbank exposures to identify emerging risks.
2. Early warning systems, stress testing, and scenario analysis
Liquidity risk modelling – stress testing and scenario analysis – can be particularly challenging and requires several critical considerations to be addressed to produce insightful results (Ong, 2014). First, is the development of realistic assumptions and scenarios. Liquidity risk models must account for the sudden and immediate impacts of shocks, along with rapid contagion effects. Key assumptions should include management responses, contagion dynamics, and second-round effects, as these significantly influence the realism of the tests. Scenario design must carefully model the behaviour of assets and liabilities, banks’ management strategies, and the transmission of shocks across the financial system. Second, effective models should capture how one bank’s actions influence the broader banking sector and potentially NBFIs. Given the increasingly significant role of NBFIs in the financial system, liquidity models must incorporate their behavioural assumptions to determine their impact on banks’ funding (Baudino et al., 2024). Third, liquidity modelling must account for second-round effects, beyond the initial shock and market responses. Modelling these effects is complex and requires robust frameworks for simulating interactions within the financial system.
Effective liquidity modelling demands access to highly granular data from a broad range of counterparties, including detailed mapping of exposures and interconnections within the banking sector to model contagion effects accurately. This means that authorities must invest in data processing and modelling capabilities to achieve more accurate assessments of funding vulnerabilities and inform strategies to enhance financial system resilience. In this regard, supervisors can deploy suptech tools to automate stress tests of banks’ resilience to various liquidity scenarios such as market shocks, sudden withdrawals, or interruptions in funding sources. Machine learning (ML) algorithms can analyse historical data to calibrate optimal assumptions and shocks to predict future liquidity shortfalls or stress periods and market reactions. Central banks can employ artificial intelligence (AI) models to run system-wide stress tests, determining how shocks could spread and identifying institutions that might trigger systemic liquidity events. Suptech can also analyse the transactional behaviour of banks and other financial institutions, identifying unusual patterns (such as increased borrowing in repo markets or sudden asset sales) that may indicate liquidity stress (Timmermans et al., n.d.). With access to high-frequency granular data and advanced modelling tools, supervisors can conduct stress tests and simulations more frequently to gain insights into systemic risks and prepare for different risk scenarios.
In the area of liquidity risk modelling, central banks are making slow but steady progress in adopting suptech. For instance, the Single Supervisory Mechanism (SSM) by the European Central Bank (ECB) has facilitated more frequent data collection (weekly and daily) to meet additional liquidity data needs (European Central Bank, 2023). The suptech tools within the SSM enable the ECB to compute the LCR with high frequency and conduct regular sector-wide bank stress tests. In the Netherlands, the De Nederlandsche Bank (DNB) is developing an experimental tool that combines monthly regulatory reporting data with daily payment systems data to define network and operational indicators, monitor liquidity flows, and estimate a daily proxy of the liquidity risk ratio of supervised institutions (Beerman et al., 2021). The tool also identifies cyclical patterns as a basis for forecasting risk indicators. The DNB’s approach promotes the use of transaction data real time gross settlement systems to identify abnormal deviations from regular cyclical patterns, which will enable central banks to intervene promptly during periods of financial stress. The modelling of second-round effects and mapping interconnectedness between banks and NBFIs remain largely unexplored and, therefore, present opportunities for suptech innovation.
3. Network analytics for interbank market monitoring and contagion risk analysis
The GFC and 2023 U.S. banking crisis demonstrated the importance of examining financial system interlinkages, because the failure or near failure of certain institutions rapidly spilled over to significant parts of the global financial system. Additionally, research has shown that interconnectedness among financial institutions can create channels for contagion. These channels can amplify or absorb shocks and create unpredictable outcomes that ultimately affect system instability (Brunetti et al., 2023).
Network analysis has increasingly emerged as an effective tool for analysing financial system interconnectedness. It enables authorities to analyse banking systems in terms of what the major triggers and channels of contagion are, and the systems’ resilience to contagion (European Central Bank, 2010). Network modelling can be used to analyse how interbank systems are affected by the heterogeneity of participants and to changes of market and bank-specific risk parameters (Halaj & Kok, 2014). Interbank networks also alleviate the challenge of information asymmetry, especially in periods of liquidity stress when it is vital to identify vulnerabilities in the market (Brunetti et al., 2023).
Suptech can be used to develop ML network models to map relationships between financial institutions and simulate contagion scenarios. This helps central banks prepare for potential liquidity crises that may spread through the financial system. The models can monitor interbank lending and liquidity transfer networks, identifying critical exposures and relationships that could propagate systemic risks. Suptech can analyse the transactional behaviour of banks and other financial institutions, identifying unusual patterns (such as increased borrowing in repo markets or sudden asset sales) that may indicate liquidity stress. The application of network models for liquidity risk assessment by financial authorities remains minimal, notable progress has been made for market surveillance (Broeders and Prenio, 2018) and AML/CFT supervision (Coelho et al., 2019).
Practical considerations for adopting suptech for bank liquidity supervision
The numerous benefits of suptech for enhanced liquidity risk oversight notwithstanding, central banks are likely to face challenges in adopting the proposed solutions. This section highlights three major challenges that would have the greatest implications for proposed use case areas in this article.
1. Resource constraints
Financial authorities considering enhancing their supervision frameworks with advanced technology should expect to undertake considerable digital transformation, which requires heavy financial investment. Several jurisdictions, particularly in emerging markets and developing economies (EMDEs), face financial constraints that hinder the acquisition, development, and deployment of advanced suptech tools (Simone di Castri et al., 2022). The hindrances are not limited to the prohibitive cost of developing the appropriate infrastructure but also apply to acquiring the appropriate human resources to use and manage the solutions. Recruiting personnel skilled in information technology (IT), data analytics, and financial supervision is challenging due to high market demand and competitive salaries. These issues are often compounded by resistance to change, mainly marked by insufficient management support for digital transformation.
To alleviate the budgetary constraints to suptech adoption, supervisors can opt to use a phased approach to structuring their suptech projects, by prioritising critical use case areas to ensure optimal resource allocation. Authorities may also consider partnerships with multilateral organisations, experienced financial authorities, and private sector organisations for financial and technical support. Skills gaps can be addressed in various ways, including implementing targeted training programs to upskill existing staff, and recruiting specialists with the necessary expertise while fostering cross-training to cover gaps. Supervisors may also explore knowledge-sharing platforms and collaborative forums to leverage expertise and best practices from other jurisdictions, such as the Cambridge Suptech Lab. Perhaps, the greatest factor to consider in conquering resource constraints is the level of senior management involvement in suptech projects. The supervision teams must prioritise securing management buy-in by demonstrating the value of suptech in enhancing bank liquidity risk oversight through regular engagements and updates, pilot projects, and case studies. Additionally, open discussions about career development will improve staff retention and guarantee quick adoption by the user teams.
2. Data quality and standardisation
Effective suptech implementation relies on high-frequency granular data, yet authorities often face issues with inconsistent data formats, poor data quality, and the absence of standardised reporting requirements. Relatedly, many authorities struggle with the integration of legacy systems, computational capacity constraints, and ensuring interoperability across different databases.
Supervisors must take steps to modernise liquidity reporting instructions to streamline regulatory data submissions and eliminate redundancies. Unified data dictionaries and regulatory data standards will ensure consistency across all reporting entities and data sources. The right data standardisation tools will support interoperability between different tools, eliminating multi-platform or database inefficiencies while reducing manual processes and allowing automated data flows. Adopting centralised data storage systems like data lakes will improve system integration and data management and analysis. Above all, financial authorities should develop or enhance their data governance frameworks to ensure quality, timeliness, and granularity of supervisory data.
3. Ethical and legal issues
The ML-enabled tools and methodologies proposed in this article may have ethical and legal implications. First, the nature of some of the data that is required for effective liquidity risk modelling may reveal banks’ consumer and counterparty identities. Second, the potential volume of data would require the use of cloud storage and computation solutions which, while efficient, present data security and confidentiality risks. As such, legal issues may arise from intended or unintended breach of data privacy laws. Furthermore, the “black box” nature of some suptech tools presents challenges arising from lack of transparency and interpretability and data bias (Financial Stability Board, 2020).
Therefore, as supervisors proceed with pioneer technologies, they need to develop comprehensive regulatory frameworks and robust guidelines to address confidentiality and cybersecurity risks. Regulatory decisions based on AI-derived analysis should be governed by policies that promote transparency in algorithmic design and improve comprehension, without the exclusion of human oversight. Authorities should build technical teams that combine supervisory and data analysis skills with the capacity to regularly audit models to identify and mitigate biases and interpret outputs effectively.
Conclusion
Supervisory technology represents a pivotal opportunity to enhance liquidity risk oversight and bolster financial stability in an increasingly complex environment. By addressing challenges related to resources, data governance, and ethics, financial authorities can unlock the full potential of suptech to modernise their supervisory frameworks.
The journey toward suptech adoption requires vision, collaboration, and innovation. By fostering partnerships, investing in human capital, and embracing technological advancements, financial authorities can navigate the challenges of liquidity risk supervision and safeguard the resilience of the global financial system.
_______________________________
For further information, we encourage you to read the State of SupTech Report 2025, access session recordings and engage in discussions on GovSpace.io.
