VoxEU Column Financial Markets

A call for liquidity stress testing and why it should not be neglected

Liquidity risks can be a primary source of bank failures. As such, there are arguments not to rely on a single metric for providing supervision. This column describes research on detailed cases of failed and near-failed institutions, which helps highlight gaps in current practices of liquidity stress testing. It also gives guidance on how to design liquidity stress tests. Deposit insurance coverage, the heterogeneity of lending commitments, distinction between different types of repos, committed facilities, and derivative transactions should receive increased attention when designing liquidity stress tests.

The recent financial crisis has shown that neglecting liquidity risks comes at substantial costs. In order to reinforce banks’ resilience to liquidity risks, the Basel Committee on Banking Supervision (BCBS) proposed the introduction of two harmonised liquidity standards:

  • The liquidity coverage ratio; and
  • The net stable funding ratio.

While the implementation of harmonised liquidity regulation across the globe is a unique and necessary step for supervision, one single metric cannot provide a complete picture of an institution’s liquidity risk profile.

Complementing the Pillar 1 standard, advanced stress tests are a useful instrument to analyse and understand an institution’s vulnerabilities. However, unlike capital for which harmonised stress tests are widespread, practices regarding liquidity stress testing still differ, and often liquidity risk is only a small component of stress tests. For their recent stress test, the European Banking Authority, for instance, only accounted for liquidity risks as an assumed increase in funding costs, as opposed to actually testing the size and quality of institutions’ liquidity buffers. Also, in the US liquidity stress tests play only a subordinated role compared to capital stress tests. This column discusses two recent Bank for International Settlements working papers (BCBS 2013a and BCBS 2013b), which present current practices, identify gaps and suggest areas of further work regarding liquidity stress testing.

While the papers draw on the academic literature, as well as banks’ and authorities’ approaches to liquidity stress testing, their core contribution is to presents and discuss detailed case studies of failed or near-failed institutions. As such, the purpose of the papers is to give guidance on how to design liquidity stress tests. In Europe, the European Banking Authority has the opportunity to facilitate the process of integrating liquidity and solvency stress tests, and to increase both the role and the quality of liquidity stress tests when addressing their mandate to issue guidelines on supervisory stress testing (Article 100 CRD 4).

How to do liquidity stress testing?

Liquidity stress testing can be considered to be a thorough analysis of an institution’s capability to compensate net cash outflows. It complements the liquidity coverage ratio as it can take into account different time horizons, different stress scenarios, and can additionally account for institution-specific and country-specific factors. Although the determination of stress test assumptions is a complex exercise, the case studies in BCBS (2013a) provide important insights into sources of liquidity stress, and should be used as input when designing liquidity stress tests. Below I will first discuss the major sources of liquidity and funding stress, and then turn to the role of liquidity buffers.

Sources of liquidity and funding stress

As credit markets froze during the recent crisis, several banks faced difficulties to execute planned sales of loans, forcing them to hold assets when they already faced funding pressures and would have liked to deleverage. These issues forced banks to increase their reliance on wholesale funding markets and, therefore, over-reliance on wholesale funding was often not the key problem, but rather a symptom of lending pipeline issues. Related to this, some banks were not prepared for the difficulties in shutting down their mortgage pipelines and, thus, experienced considerably lower inflows than anticipated. Simultaneously, the capital position of these institutions worsened given the unplanned asset on-boarding.

  • A first insight is that banks cannot rely too heavily on inflows and, therefore, when designing liquidity stress tests, inflow rates should be calibrated conservatively.

Turning to outflows, the cases of two US commercial banks, discussed in BCBS (2013a), show that uninsured deposit outflows were one of the drivers of liquidity stress, while one large international bank experienced deposit outflows of 11% over one month, with 8.5% within a single week.

  • As such, the case studies clearly show that uninsured deposit outflows are not a negligible liquidity risk.
  • Another important insight from the case studies is the heterogeneity of lending commitments.

While commitments to corporate borrowers, for instance, were one of the smallest sources of liquidity stress during the recent financial crisis, commitments to asset-backed commercial paper conduits and other capital market instruments significantly affected banks’ liquidity positions.

The example of a European bank shows that while margin calls and pre-funding of FX swaps reduced the bank's liquidity position, these factors amounted to only 8% of the firm's liquidity gap and, therefore, were a contributing as opposed to a major factor. Similarly, data for a US investment bank suggest that collateral movements due to derivative assignments amounted to only 10% of the firm’s outflows, and liquidity risks stemming from derivative transactions should not be calibrated overly conservative.

• Other major funding issues were caused by run-off in prime brokerage balances ($46.7 billion during a two-week period of the 2007–09 financial crisis for one bank), claims from secured lending counterparties, as well as liquidity stress related to the use of clearing and settlement services.

The role of liquidity buffers

To cope with large net cash outflows, banks hold liquidity buffers. Zooming into these buffers, BCBS(2013a) shows that one of the European institutions carried a buffer of €2.1 billion while facing asset-backed commercial paper-related outflows of roughly €20 billion. In line with this, three other case studies show that banks’ liquidity buffers were simply too small relative to the stress they experienced during the crisis (as opposed to banks experiencing difficulty in liquidating those assets).

Several banks, however, also had difficulties to access their liquidity buffers. Most interesting is the case of one US firm which developed a number of creative approaches allowing it to include assets provided to clearing and settlement banks in its liquidity buffer, even though these funds could not be accessed during regular business hours. Another source of liquidity risk was the so-called ‘lemons problem’ -- difficult to value, complex products suffered from higher discounts, and have proven to be a less reliable source of liquidity, especially when they were not traded in deep and active markets. Clearly, this stresses the importance of well-defined operational requirements when designing liquidity stress tests.

Finally, the case studies suggest that several firms were able to raise funds through repo with central counterparties (CCPs), but that there are reasons to view firms' continued access to bilateral repo markets with scepticism. For one institution, CCP-intermediated repo served as a material source of funding even during the most significant period of stress. A bank was able to raise $24 billion through general collateral finance repo during its peak liquidity stress and similarly, one failed US institution experienced no material change in triparty repo haircuts or financing volumes until one week prior to failure. As such, this may suggest taking into account the repo-ability of an asset with a CCP as a criterion for favourable treatment in liquidity stress tests.

The table below shows potential stress parameters that banks and supervisors may consider when designing liquidity stress tests. It draws on the evidence presented in BCBS (2013a), BCBS (2013b), the stress tests discussed in Schmieder et al. (2012), but also the rationale behind the liquidity coverage ratio assumptions. Naturally, the parameters are only suggestions and there are many factors that can justify large deviations. Especially when designing tailored stress tests for individual institutions, stress test parameters might be considerably different, depending on other factors, such as the institution’s business model, relative rating, or solvency.

Table 1. Potential stress parameters

More general remarks

Apart from these more specific issues, BCBS (2013a) also points to a number of more general observations to bear in mind when developing liquidity stress tests. Ideally, competent authorities apply both bottom-up and top-down approaches to capture second-round and systemic effects. Similarly, it is also recommendable for banks to take into account second-round effects, and especially large banks should account for the impact of their actions on the banking system as a whole.

Regarding central bank funding, the paper discusses arguments in favour and against factoring in the lender of last resort function to supervisors’ liquidity stress tests. On balance, the paper concludes that liquidity stress testing assumptions should limit the role of central banks to standard and other already existing monetary policy operations while the assumptions of more expansive central bank support would give wrong incentives with regard to banks’ risk taking and reliance on central bank liquidity.

Conclusions and final remarks

BCBS (2013a) once again confirms that liquidity risk is a complex and diverse matter. At the same time, however, the case studies clearly point to a number of patterns. Especially the importance of deposit insurance coverage, the difficulties faced by banks to stop leveraged and residential mortgage loan pipelines, the relevance of operational requirements for the effectiveness of liquidity buffers, as well as the important distinction between different types of repos, committed facilities, and derivative transactions, should receive increased attention when designing liquidity stress tests.

Liquidity risks can be a primary source of bank failures. As such, there are strong arguments not to rely on a single metric but to conduct serious Pillar 2 liquidity supervision with supervisory and firms’ own stand-alone liquidity stress tests as well as stress tests combining liquidity and solvency risks being the quantitative fundament of such a process.

Editor's note: The views expressed in this column are those of the author and do not necessarily reflect those of De Nederlandsche Bank.


BCBS (2013a), “Liquidity stress testing: a survey of theory, empirics and current industry and supervisory practices”, BIS Working Paper No. 24, October.

BCBS (2013b), “Literature review of factors relating to liquidity stress – extended version”, BIS Working Paper No. 25, October.

CEBS (2009), “Guidelines on Liquidity Buffers & Survival Periods”, 9 December.

CEBS (2010), “CEBS Guidelines on Stress Testing (GL32)”, 26 August.

DNB (2013), “Guidance for banks when stress testing for liquidity risk”, DNBulletin 7 November.

FRB (2013), “2014 Supervisory Scenarios for Annual Stress Tests Required under the Dodd-Frank Act Stress Testing Rules and the Capital Plan Rule”, Board of Governors of the Federal Reserve System, 1 November.

Schmieder C, H Hesse, B Neudorfer, C Puhr and S Schmitz (2012), “Next Generation System-Wide Liquidity Stress Testing”, IMF Working Paper WP/12/3.

2,205 Reads