More Liquidity Stress Testing: Is It Enough to Prevent Future Crises?
The dust is only just settling after the recent turmoil propagated by the rapid-fire failures of Silvergate Bank, Silicon Valley Bank (SVB), and Signature Bank. Recall that the Federal Deposit Insurance Corporation (FDIC) took over SVB on March 14, 2023, and then two days later announced it would also guarantee its uninsured deposits. The crisis was further marked by the US Federal Reserve and the FDIC taking extraordinary actions to assure depositors that their funds are still safe. Similarly, European regulators had to deal with the near failure of Credit Suisse and subsequent intervention by the Swiss government and Swiss Financial Market Supervisory Authority (FINMA) to broker its fire sale to rival UBS.
Historically bank runs occurred over days. Today, they happen in just a few hours.
These failures gave market participants extraordinary reasons for concern and led to swift regulatory actions. Historically, runs on banks occurred over days, but in the digital age they happen in a matter of hours, rapidly cascading across similar institutions, leading to concerns that other banks, many of whom were otherwise healthy and had strong balance sheets, may soon follow suit.
Key Changes That Strengthen Liquidity Risk Management
As regulators and industry ruminate on recent events, surprising shortcomings in the liquidity-risk management of these firms have emerged as underpinning weaknesses, crystalizing three key areas for improvement across the broader industry.
- Regulatory mandated 14- and 30-day horizons for stress testing under Regulation YY lack sufficient granularity.
- Liquidity stress testing (LST) modeling assumptions need to be more sensitive to business dynamics and anchored by granular data cohorts.
- Run-off rates like those for unsecured wholesale deposits and non-high-quality liquid assets (HQLA) secured funding, should be more severe.
A new focus on liquidity stress testing
The liquidity coverage ratio (LCR) is a regulatory one-size-fits-all risk model with a universal regulatory limit. In contrast, LST covers multiple models for regulatory and business requirements with limits. Considering current market conditions, liquidity vulnerabilities are heighted for lenders and depositors alike. So, changes in related risk-measures and regulations are both more likely to materialize and have a material impact. Changes could include:
- Considering original maturities as a liquidity-risk-sensitivity criteria
- Compounding LST rollover/run-off rates based on original maturities
- Adding a short-term unsecured wholesale-funding regulatory limit
- Requiring the public disclosure of held-to-maturity (HTM) positions marked-to-market unrealized P&L (net of hedges)
- Monitoring the interest rate risk in the banking book (IRRBB) changes
- Including regulatory capital add-ons
Assessing LCR and LST, at this juncture
Both processes should be based on the same data, with the same controls and governance; and, as previously mentioned, the frequencies of measure should be greater than the regulatory frequency. Specifically, time horizons need to be more than 14 and 30 days.
In addition, the LST capability must include data not used in LCR like uncommitted credit facilities. It should also:
- Have more extreme scenarios in its run-off rate severity
- Be proactive to business requirements instead of reactive to regulatory requirements
- Have stress-scenario results expanded to include net-funding requirements and liquidity-risk grouping-policy-limits compliance
- Show and report sources and uses of liquidity, with net daily and cumulative outputs
Deeper scrutiny into data, models, and calculations — end-to-end
Undoubtedly the recent turmoil will prompt regulators to pay more attention to the stress-testing methodologies and processes that financial institutions use, increasing the pressure on them to show their calculations and associated concrete data and demonstrate exactly how those calculations and data are governed by internal controls. Therefore, banks’ continued reliance on simple monthly Excel-based modeling will no longer be sufficient to withstand heightened internal and external scrutiny.
Likewise, the increased scope of calculations and analytics will require a comprehensive review of strategic data sourcing and architecture. Maintaining tight data lineage to prove out accuracy will be mission critical, especially considering the volume of change.
Equally paramount is the need for much deeper and comprehensive transparency into results, pared with the ability to dissect the data at varying levels of granularity and scenario treatments. Business users will need to quickly run different scenarios on source data and simulated data sets to assess current and future risks and opportunities.
Prepare Now, Before The Next Shoe Drops
There are key measures that financial institutions can take right now to secure their balance sheets against market turbulence. However, achieving a strategic, futureproofed, and flexible foundation for liquidity risk-management likely means sunsetting piecemeal, manual-process-intensive systems/solutions, and/or black-boxreliant approaches.
12 factors to consider for a robust, data-driven liquidity risk management solution
The hallmark of an effective LST solution is that it strongly contributes to a bank’s confidence in its liquidity risk management, futureproofs it against market volatility and regulatory change, and fosters trust between the bank and its regulators. Solution characteristics can be summarized as follows. It:
-
1. Easily Integrates with existing FR 2052a reporting and other external data
-
2.Provides accuracy and reconcilability via an extensible regulatory data dictionary-based architecture
-
3.Enables the modelling of idiosyncratic scenarios on the same, complete, and reconciled data used for FR 2052a reporting
-
4.Demonstrates clear lineage between source data and calculation results by efficiently utilizing the same data set to run optimization, stress testing, and what-if analysis
-
5.Is highly performant on large data volumes for rapid analysis, re-runs, and timely reporting
-
6.Gives business users a transparent view of and control over the applied business rules and logic
-
7.Allows the business to flexibly add new scenarios as regulatory and business needs change
-
8.Manages the stress-testing process in a controlled environment with granular permissions and roles. It should proactively support business requirements rather than react to regulatory requirements
-
9.Expands stress scenario results to include net funding requirements and liquidity-risk grouping policy-limits compliance
-
10. Shows and reports sources and uses of liquidity with net daily and cumulative outputs
-
11. Integrates solutions across the trading and banking books, providing a clear path to a strategic approach for both credit and market risk
-
12. Enriches the target operating model with embedded controls and data governance that facilitate day-to-day liquidity-risk management; and enables successful internal audit and regulatory examinations
From a system perspective, this anticipated approach to LST corresponds to a new set of requirements including the management of comprehensive and granular datasets, advanced processing capabilities that cater to more frequent executions and granular monitoring horizons, and reporting capabilities that deliver targeted insights for decision making.
Our Nasdaq AxiomSL Liquidity Stress Testing Solution on the Nasdaq AxiomSL ControllerView® platform enables you to combine FR 2052a liquidity data with other sources to model various scenarios and much more transparently and flexibly manage your liquidity risk. In addition to addressing US-related situations, our holistic solutioning approach meets your liquidity risk management needs globally, while taking care of jurisdictional interpretations/nuances.
