Skip to main content

Revisiting Intraday Liquidity in a Volatile Economy

The high cost of funds in an inflationary economy needs a modern system to manage intraday liquidity risk of banks.

Overview

 

The decade of low cost funds has led to intraday liquidity being a highly neglected area and only a compliance exercise for many banks. However, the recent macroeconomic developments leading to the high cost of funds and shrinking liquidity made it necessary to revisit the intraday liquidity forecasting to optimise the intraday fund usage and identify emerging liquidity crisis. In this article we provide our view on how banks can modernise intraday liquidity risk management with the help of better data availability, modern technologies and leveraging a better modelling technique "Nowcasting".

 

Introduction

 

The Basel Committee released the final version of the Intraday Liquidity Monitoring in April 2013. However, in the past 10 years its importance was reduced to merely a regulatory compliance requirement for banks due to the low/negative cost of funds. The situation has changed drastically in the last few months as geopolitical developments, higher inflation, and shrinking global liquidity have made the cost of funds rise in volatile spurts. Given the current market circumstances, with rising and volatile interest rates, there are two major impacts of immature liquidity forecasting:

  • Missing out the opportunity cost of fund deployment and 
  • Unable to identify emergent crisis situations.

An effective intraday liquidity forecasting approach allows a bank to anticipate major sources of inflows and deploy the funds into the most remunerative funds usage. This activity also signals to the money market counterparties the health of the institution and proactive interaction with the central bank in the event of unexpected liquidity shortfalls. The importance of this forecasting activity is evident, as established and long-standing institutions around the world have recently faced rapidly evolving liquidity runs due to ineffective cash flow projections and its correlation with emerging financial market sentiments.

Conventional liquidity forecasting models, relying on siloed data and a disintegrated governance framework, are not able to respond to the need for reliable real-time liquidity forecasting and meet the market obligations that undergo rapid change during volatile market conditions. 

Technological advancements, payment guidelines, and increased relevance of optimised liquidity management are driving innovative developments in intraday liquidity forecasting at the leading banks worldwide. This is a lesson that other banks may use to improve their capability to manage intraday liquidity risks and ensure they have robust tools to respond to emerging liquidity traps with liquidity allocation, repo operations, and emergency funding support.

Developing a new approach to respond to the current market situation

 

Over the last 15 years most central banks have developed real-time economic forecasting approaches for tracking GDP growth patterns with an approach known as Nowcasting. This approach leverages unconventional data sets to augment existing data and leverage this insight for real-time updates to the GDP forecast. We observe that the large financial institutions are adapting this approach for their own liquidity forecasting and also for their corporate customers in order to differentiate in a competitive market.

Intraday liquidity Nowcasting is a class of NextGen model that supports critical decision-making strategy by providing a real-time prediction of cash flow more accurately than the current models. The model aggregates signals from payments data, financial markets, and other customer-specific signals to update the forecasts in an automated manner. By integrating real-time data, including numerous scenarios, and utilising sophisticated machine learning approaches, banks can enhance and develop their intraday liquidity Nowcasting model. 

The model development typically involves standard steps such as identification of right data attributes, data management, modelling, monitoring, and decision making. However, the important aspect is to implement an unsupervised learning approach for processing the signals as they emerge in the various streams of information, such as payments, customer sentiment, etc. Some key business and technology elements which must be taken into account while developing a liquidity Nowcasting model are:

  • Consumer behaviour, financial market interactions, macroeconomic data, and internal bank data combined with enriched payment transaction data from ISO 20022 messages. 
  • ISO 20022 payment messages are key enablers which provide transaction data such as payer, payee, and purpose of the payment in a more reliable and detailed manner. The unstructured additional details are unlimited and can provide additional information in understanding, for example, the nature of the transaction and the potential effects on the liquidity situation. The improved quality of data provides reliable real-time input to a more sophisticated intraday liquidity forecasting model.
  • Machine learning approaches to financial time series analysis in combination with large language models are being applied for real-time cash forecasting for corporate customers. Several large banks in the world have begun developing these models. Therefore, small/mid-size banks should invest more to stay in the race.
  • The availability of cloud-based ML Ops frameworks can be leveraged to enable rapid development and deployment of the model.
  • And probably the most important enabler is setting up a cross functional team consisting of business experts, data scientists, and tech experts in order to address this multi-dimensional problem and deliver rapid business outcomes.

Structured data, such as those from financial and payment transactions, can provide critical real-time information for assessing aggregate consumption and economic activity. …For Italy, some gains in forecast accuracy have been reported when information from highly aggregated but large value payments (i.e. TARGET2) has been included in GDP nowcasting models.

 

Overview of an intraday liquidity nowcasting model 

The entire process of data grouping, selection, calibration, and validation of the model is automated once the parameters for the forecasting and the necessary data have been identified. Also, the computations get more precise over time as more data with the appropriate level of granularity becomes available. 

The learning algorithm automatically decides if the existing model is still the most suitable or whether a different model might produce superior outcomes before making each new forecast. The accuracy of predictions increases with time as the data foundation for calculations is continuously being updated with real-time data sets from multiple sources.

The final output from the intraday Nowcasting model is represented in a dashboard with charts, graphs, and other visualisations for hourly and daily (up to next 30 days) cash flow projections. The dashboard will typically include cash inflow and outflow from various payment types (such as remittances, trade finance, securities settlements, direct debits, card payments, etc.) and also provide a view according to segment (retail, corporate, and financial institutions).

The framework for overall liquidity management is increasingly being combined with intraday liquidity management processes. Banks are approaching liquidity management more comprehensively, including both short-term and long-term cash requirements, and creating integrated strategies to control liquidity risks.

Subsequently, the intraday liquidity Nowcasting model can be integrated with the existing liquidity management framework used in treasury and cash management within a bank or any organisation.

Conclusion

Most banks found it difficult to put intraday liquidity ideas into practice, especially when it comes to data integration and quality. Also, the challenge has been to move beyond pure regulatory compliance and provide business insights.

Today, the majority of liquidity management concerns are still not fully addressed. A holistic view of intraday liquidity across clearing, correspondent, and central bank accounts is one of the challenges that still exists due to inconsistent and incomplete transaction data when real-time “always-on” visibility on liquidity sources is much needed.

In light of the evolving banking landscape, it may be necessary to revisit BCBS 248 principles in order to solve the problems brought by current market conditions and give banks more choice on how to handle new risks as the banking industry continues to interact with counterparties beyond banking (such as open APIs, fintech players, etc.).

However, there is hope in the new algorithms and technology developments which allow parsing of large data sets (both structured and unstructured) and consume them in powerful AI/ML models for drawing conclusions automatically. This allows deployment of Nowcasting algorithms for intraday liquidity forecasting which can start delivering business and risk insights in a fairly short timeframe. Hence, this is an initiative which shall deliver business value if the banks re-architect the solution framework while keeping in mind the business drivers and technology flexibility in today’s volatile economy.

Abbreviations used:
AI - Artificial intelligence
APIs- Application Programming Interfaces
BCBS 248- Basel committee on Banking supervision, BCBS 248 is a document published by the Basel Committee on Banking Supervision in April 2013 in the Liquidity Risk category
GDP- Gross domestic Product
ISO 20022- International organization for standardizations. ISO 20022 is a standard for electronic data interchange between financial institutions
ML- Machine Learning
Ops- Operations

Key contacts

Anshuman Choudhary

Partner in Risk Advisory, Regulatory and Legal Support

Anshuman is a Technology Partner within the Risk Advisory department at Deloitte Belgium and has over 22 years of professional experience in the Financial Services Sector. He has joined Deloitte recently from Cognizant Technology Solutions where he had worked for the last 15 years. Since August 1998, he has been working on various assignments in the financial sector, in engagements ranging from risk audits (credit, market, & non-financial risks) to large scale system implementation pertaining to core banking platforms and risk data-warehouses. For Deloitte Belgium, Anshuman is developing offerings in four areas a) Credit and Liquidity risk management automation, b) Financial Crime process orchestration, c) Liquidity Risk Automation for Asset Managers and d) Model Risk Management. Anshuman is also the Technology Partner for a number of FSI customers in Belgium responsible for driving technology projects with customers.

Deepak Sharma

Manager, Financial Industry Risk and Regulatory

Deepak has more than 14 years of global experience in Banking and Financial services (BFS) domain across large financial institutions. He has subject matter expertise in Domestic & International Payments, Cash Management, KYC, Transaction monitoring and Digital Banking area. His recent experience include implementing ISO 20022 Payments for different market Infrastructures, Sepa Direct Debit APIs and Data Assurance for Investments data. Deepak has specialized in leading large Banking and Regulatory projects and defining global Business, Operations & IT roadmap for the clients. His International experience with top BFS clients has helped him gaining strong stakeholder management, communication and analytical skills while understanding new ways of working in a geographically disperse team.

Wangting Chen

Senior Consultant, Regulatory and Financial Crime

Wangting is a senior consultant within the Regulatory and Financial Crime team of the Risk Advisory department at Deloitte Belgium. Her expertise include risk management and anti-money laundering developed during her work with international banks in China, Portugal, Belgium and Netherlands. Wangting is passionate about helping banks and other financial institutions in compliance to risk management requirements to create added value for their customers.