Introduction

• The global liquidity crisis, which lasted from approximately August 2007 to the end of 2008, ushered in the broader financial crisis and highlighted the importance of prudent management of liquidity risk. The liquidity stress test provides the critical underpinning to a bank’s liquidity risk management framework by determining the amount of liquidity that must be held in order to ensure the institution can meet financial obligations under stressed conditions. A robust liquidity stress test is based on a projection of cash flows arising from assets, liabilities, and other off-balance sheet items under a variety of systemic and idiosyncratic scenarios that can occur over varying time horizons. The results of the liquidity stress test provide the foundation for setting the bank’s liquidity risk appetite, establishing appropriate limits and targets, and developing an effective contingency funding plan.

• Two of these emerging trends n liquidity stress testing programs are –
  • First, the industry should improve the level of integration and consistency between the liquidity stress test, the capital stress test, and, more broadly, risk measurement and monitoring, performance measurement, and regulatory reporting.
  • Second, banks should invest in achieving a sustainable technology infrastructure that ensures liquidity stress testing is performed in an efficient and controlled manner.

Measuring Contingent Liquidity Requirements

• It is important to clearly define what is meant by “liquidity” for liquidity stress testing purposes. Within this context, liquidity refers to funding liquidity risk – the risk that the institution will not have adequate capacity to fund its obligations without incurring unacceptable economic losses. Assessing asset liquidity- the risk of incurring losses due to difficulty converting assets into cash-while not the objective of the liquidity stress test, must also be carefully considered in the liquidity stress test since it can also impact the amount of funding that can be made available from the sale of assets.

• As a source of funding liquidity, businesses, including financial institutions, utilize liquidity for four purposes:
  • Operational,
  • Restricted,
  • Contingent, and
  • Strategic.

• Operational liquidity represents the cash that is needed to fund the business on a daily basis, and it is required to ensure orderly clearing of payment transactions. Depending on the nature of the institution’s business, operating cash needs might be quite volatile and, as a result, a cushion is added to account for the unpredictability of daily settlements and the excessive system and management effort that would be required to manage cash to its theoretical minimum. Operational liquidity must be maintained to ensure the institution’s operations and is therefore unavailable to meet financial obligations under a liquidity stress test.
• Restricted liquidity represents liquid assets that are available to be used only for specifically defined purposes. For example, a bank may be required to collateralize certain wholesale borrowings. Restricted liquidity is unavailable to meet general financial obligations under a liquidity stress test, but should be applied to any assumed outflows which they support.

• Contingent liquidity represents the liquidity that is available to meet general financial obligations under a stress scenario. This liquidity is available in the form of the institution’s liquid asset buffer, which comprises access to financial assets that are of very high quality and can be easily converted into cash without any real loss of market value. Measuring required contingent liquidity to cover stressed cash outflows is the principal objective of the liquidity stress test.
• Strategic liquidity represents the cash that is held by the institution to meet future business needs that may arise outside the course of normal operations, but it is not primarily aimed at supporting the bank during times of stress. For example, strategic liquidity may be held to fund future acquisitions or capital expenditure programs. Strategic liquidity may be redirected to meet contingent liquidity requirement needs. As a pragmatic matter, this will likely be feasible only if such liquidity is present via holdings of highly liquid assets.
• This liquidity taxonomy is illustrated in this figure

Estimating Continegent Liquidity

• To measure the amount of required contingent liquidity, then the institution must construct a cash flow model that accurately and precisely measures the following components:

• Liquid asset buffer – The liquid asset buffer represents the contingent liquidity that is currently in place.
  • The liquidity-generating capacity of securities included in the liquid asset buffer should remain intact even in periods of severe idiosyncratic and market stress.
  • The fundamental characteristics of liquid asset buffer securities should include low credit and market risk, ease and certainty of valuation, trading in an active and sizable market, and low concentration of buyers and sellers.
  • The liquid asset buffer should also meet operational requirements that ensure the liquidity is under the control of the central treasury area of the entity undergoing the stress test.

• Stressed outflows – Stressed outflows are those assumed to occur under stress scenarios.
  • Stressed outflows may result from the need to prematurely settle non-contractual maturity obligations as well as the inability to refund contractual maturity obligations that under normal circumstances could be rolled over.
• The types of outflows to be modeled typically fall into the categories of retail deposit outflows, unsecured wholesale funding outflows, secured funding runoff, derivative transaction funding, loss of funding on asset-backed issuances, and drawdown of credit and liquidity facilities.

• Stressed inflows – Stressed inflows are assumed to partially offset the stressed outflows.
  • Inflows may include secured funding transaction maturities, loan repayments from customers, and drawdowns on liquidity facilities available to the institution.
  • Depending on the assumptions used in a particular stress scenario, the level of inflows may be reduced or limited by market conditions.

• The stressed liquid asset buffer indicates the adequacy of the current liquid asset buffer given the stress scenario assumptions. Using the above three, it can be estimated as liquid asset buffer — stressed outflows + stressed inflows

Design Issues of The Model

1. Organizational Scope – The consolidated stress test should be the pillar of any liquidity risk framework.

• However, an institution may determine there is a need to conduct stress testing on subsidiary entities within the organization. The organizational levels at which a bank may stress liquidity include the parent, subsidiary legal entities, lines of business, service business units, and shared service centers. Each of these cases may be addressed through a separate liquidity stress test, where necessary.
• For less material entities or those entities where risk is assessed to be manageable, less complex entity-level liquidity risk reporting might be sufficient. As a general rule, the institution should consider the organizational level at which (a) liquidity is commingled, and (b) liquidity oversight has management accountability. Combinations of legal entities and operating units having both of these characteristics will provide the building blocks of the enterprise-level liquidity stress test.

• Liquidity transfer restrictions – Liquidity may be trapped in certain legal entities, potentially creating a distorted view of the consolidated liquidity position of the institution. For example, foreign exchange controls may inhibit the conversion of foreign currency in off-shore legal entities. The bank should assess the impact of such restrictions on enterprise level liquidity, considering not only a normal operating environment but stressed conditions as well. The existence of liquidity transfer restrictions does not necessarily give rise to the need for an additional stress test where it can be demonstrated that a subsidiary would not be required to upstream cash to the parent.
• Currency – While the liquidity stress test should be performed in the currency of the entity being tested (the home country for the consolidated test), careful consideration should be taken for the liquidity impact of currency conversion requirements. For example, less established offshore subsidiaries or branches sometimes carry a significant currency mismatch, and the settlement time frame for the home country parent to swap fund an unanticipated outflow may prove problematic in a crisis.
• Regulatory jurisdiction – For institutions operating in multiple foreign jurisdictions under various regulatory oversight regimes, the need to conduct individual stress tests for foreign subsidiaries or groups may arise.

2. Planning Horizon – The planning horizon of the liquidity stress test should be at least twelve months.

• The bank may choose to project cash flows beyond twelve months; however, longer-term projections may be subject to significant forecast error depending on the time horizon of the baseline balance sheet and income statement budgeting performed as part of the strategic planning process.

• One circumstance in which the bank may choose to forecast beyond twelve months is the case where a survival horizon is calculated under the stress test. For banks with ample liquidity, the survival horizon may extend well beyond this period; some banks have a survival horizon that may extend as far out as two years, although the extent of the modeled stress will abate beyond the extreme level of severity assumed in the very short term.
• The frequency of cash flow measurement within the overall time horizon must also be determined. The decision to estimate daily, weekly, or monthly cash flows should balance the benefits of improved precision against the reduced forecasting accuracy beyond a certain time frame. Stress models that forecast daily over a short time frame (e.g., one month) and transition to weekly or monthly cash flows for the remaining time horizon are likely to provide the best balance. The need to forecast daily during the initial stage of the stress test is recommended not only as a result of the relatively higher predictability of these cash flows, but also because, as was seen during the financial crisis, the most critical period of stress for the institution may in fact occur during those first few days.

Testing Techniques

There are three general approaches to performing a liquidity stress test –

• Historical statistical techniques,such as cash flow at risk (CFaR), model a historical pro forma cash flow based on the observed cash flow volatility of the institution.
• Deterministic models,are used to model the liquidity impact of a forward-looking or historical-based scenario that has been developed by the institution.
• Monte Carlo simulation relies on simulation modeling and can be used to assess liquidity risk by stress testing specified variables over a future time frame.

• Stochastic techniques that rely on observations of historical volatility of cash flow variables, whether using historical statistical models such as CFaR or Monte Carlo simulation techniques that rely on historical observations of volatility, are less popular after the financial crisis. An additional challenge of stochastic approaches is their limited ability to accurately predict the management countermeasures that would occur during a liquidity crisis event.

• By its nature, liquidity stress is an extreme “tail event”, and deterministic scenarios, despite their reliance on many assumptions that are derived through expert judgment, are viewed by regulators and most financial institutions as the most effective tool for assessing liquidity risk.

Baseline Scenario

• The starting point for building the liquidity stress test is the baseline balance sheet funding and liquidity plan. As a banking organization builds out its liquidity stress test framework, it is sometimes necessary to enhance the structure of the baseline plan as well to ensure that the base case is consistently structured and at the same level of detail as the stress scenarios. It is also advisable to house the baseline analysis in the same reporting and analysis platform as the liquidity stress test. The objective is to ensure that the institution can gauge the severity of each stress scenario by making a valid comparison to the baseline forecast.

Scenario Development

• By its very nature, liquidity failure is a high-impact, low frequency event. Fortunately, only a handful of large financial institutions have collapsed due to insufficient liquidity. Unfortunately for this very reason, there is little data upon which to build reliable, predictive models that can accurately estimate the minimum level of liquidity an institution can expect to maintain within a confidence interval. As a result of these limitations and the highly complex, interconnected nature of liquidity behavior, the industry approach to performing a liquidity stress test is to develop a set of discrete, deterministic scenarios. While liquidity stress test scenarios have become somewhat standardized across the banking industry since the financial crisis, it is important that each financial institution carefully consider its unique or idiosyncratic material risks when building out its scenario framework. There are two general types of liquidity stress scenarios – historical scenarios and forward-looking (hypothetical) scenarios.

1. Historical Scenarios– Historical scenarios are based on actual liquidity failures and attempt to translate those events to the financial institution performing the stress test. The failures of WaMu and Northern Rock in 2008 are common reference events.

• The advantage of historical scenarios is that they are empirically based.

• The disadvantage of this approach is that few such failures have actually occurred; and for the ones that have taken place, very limited data are available. Additionally, future business conditions may cause new and unanticipated liquidity events, creating a potential blind spot for management.

2. Hypothetical Scenarios– Hypothetical scenarios are based on a forward-looking view in which the financial institution experiences severe liquidity stress. Banks typically develop multiple scenarios. Liquidity stress scenarios should exhibit the following characteristics:

a) Distinguish between systemic and idiosyncratic risk – Some liquidity stress impacts are the result of systemic stress, such as a reduction in the market liquidity of securities, while other impacts are the result of bank-only stress, such as a deposit run. Banks should develop at least one scenario for each of the cases of systemic, idiosyncratic, and combined idiosyncratic and systemic in order to capture these varying impacts.

b). Distinguish between levels of severity – Assuming graduating levels of severity, for example, by developing adverse and severely adverse variations of the idiosyncratic scenario, enables the institution to broaden its view of liquidity risk and applicable limits.

c) Clearly define the scenarios – The bank must establish a specific, detailed description of the business and market events associated with each scenario in order to provide a comprehensive view into the specific conditions the institution is experiencing. Scenario descriptions typically include the following:

• The general level of stress (e.g., high) of market, economic, and credit conditions
• Conditions of wholesale secured and unsecured funding Markets
• Changes in counterparty haircut requirements by collateral Type
• Liquidity impacts on securities in the liquidity buffer and other assets in the event of sale
• Details of credit grade downgrades
• Deposit runoff assumptions by product and customer type, and with consideration to other factors such as insurance Coverage
• Description of impacts on specific counterparty Relationships
• Rating trigger impacts on derivative margin and collateral Calls
• Impact of regulatory actions or limit breaches in foreign Jurisdictions
• Assumed drawdowns on unfunded credit and liquidity Facilities
• Assumed debt calls and buybacks

d) Consider more holistic approaches to scenario development – Standard industry liquidity stress test scenarios, including those required by regulators under the Basel III liquidity coverage ratio (LCR), are highly prescriptive. The institution should also consider scenarios based on broader economic and business impacts. Doing so will ensure the bank is considering systemic, interdependent risk behavior rather than simply developing isolated liquidity assumptions.

• In addition to assumption-based hypothetical scenarios, the bank may also perform a reverse liquidity stress test. The objective of this is to determine which conditions would need to exist, given the bank’s current liquidity level, to cause its current business plan to become unviable.

• To construct such a scenario will require determination of which factors will have the most significant impact on liquidity and stressing these assumptions to the institution’s destruction.
• Developing such a reverse stress test scenario, while simple in theory, presents a number of problems as there are a large number of factors which could combine to destroy the institution. It can also be difficult to develop a destruction scenario if the bank is highly liquid without making fantastic, apocalyptic assumptions. As a result, reverse stress testing is not a universally performed exercise among financial institutions.

Development of Assumptions

• Liquidity stress testing is built on hypothetical and historical scenarios, and as a result is highly dependent on the validity of assumptions in generating meaningful results. For many key assumptions there is limited historical or market data to draw upon in building a fact base. Nevertheless, applying segmentation frameworks that enable differentiation of assumptions across varying levels of cash flow risk enhances the rigor of the liquidity stress test. Generally, in developing liquidity stress testing assumptions the institution should follow these guidelines:

1) Qualitatively assess the expected liquidity behavior for each type of cash flow to determine where there is significant liquidity risk.

2) Determine the appropriate level of segmentation for each type of risk based on an assessment of behavioral differences, bearing in mind any limitations in ongoing data availability.

3) Qualitatively assess and order by rank varying levels of liquidity risk for each segmentation factor, potentially utilizing a scoring system.

4) Develop quantitative modeling assumptions based on any historical data available, such as experiences during the financial crisis, or available from other sources such as peer benchmarking.

5) Develop matrices of relative modeling assumptions based on scored risk levels and baseline historical data.

6) Adjust assumption matrices as appropriate for each stress scenario, for example, reflecting differences in relative overall severity or assumptions concerning idiosyncratic or systemic risk.

• The following assumptions can have an outsized impact on the results of the stress test, and should be considered carefully in developing the model:

1. Investment portfolio haircuts– For systemic stress scenarios, it is assumed that haircuts will widen on securities as was observed during the crisis. The model should include varying haircut assumptions for each security type where liquidity characteristics differ, for example, differentiating between agency mortgage-backed securities and non-agency mortgage- backed securities. The model should also include expected haircut differences between secured financing channels used by the institution, for example, Federal Home Loan Bank funding, and repo facilities. The starting point for developing liquidity haircuts is a review of current market conditions (assuming such conditions are normal), and comparing these to what the bank experienced during the financial crisis. If the bank does not have such data available, it will need to be obtained through peer comparisons where possible.
2. Deposit outflows – Deposit runoff is, for most institutions, the most significant threat to liquidity and the most important behavioral dynamic to model. For the typical, heavily deposit funded bank, liquidity stress test models built on simplistic assumptions concerning deposit behavior will most likely yield meaningless results, even if other aspects of the model have been calibrated rigorously. Unfortunately, there is a scarcity of historical data to rely upon in developing deposit runoff assumptions. While the runs that occurred during the crisis, particularly those at WaMu and Northern Rock, provide useful reference points, the institution should build a set of detailed deposit runoff assumptions based on a behavioral segmentation framework that captures differences in stressed deposit behavior. This table summarizes behavioral differences typically observed in deposit portfolios.

3. Unsecured wholesale funding– Availability of unsecured wholesale funding is generally assumed to be heavily reduced in a stress scenario, particularly under idiosyncratic stress. The bank should review each funding channel to differentiate by key liquidity factors, most significantly overnight versus term funding. There is likely to be little historical data available assuming the institution has not experienced a significant stress event. Banks typically apply highly conservative assumptions when reflecting on the drastic impact that a stress event is likely to have on wholesale funding availability, particularly term funding.

4. Collateral requirements– Collateral requirements should be expected to increase during a stress scenario as a result of both valuation impacts on existing collateral as well as increased collateral levels required as a result of changes in derivative positions. How the institution develops assumptions for collateral call levels (as opposed to collateral valuation impacts, which should align to unsecured wholesale funding models) will depend on the level of detail required. The institution may choose to review its historical collateral call levels, particularly during times of stress, and select the most significant liquidity requirement experienced during a historical period. A more detailed approach would be to model each position independently.

5. Other contingent liabilities– The model should address each material source of contingent liquidity outflow, including drawdowns of customer credit lines, liquidity facilities, letters of credit, trade financing arrangements, securitization facility runoff, and other contractual arrangements. Where possible, the institution should review the behavior of such contingent liabilities during the financial crisis. If historical data is not available, conservative assumptions are appropriate. Non-contractual commitments must also be incorporated into the model. Particularly for lower stress scenarios, the institution will still seek to maintain reputational strength and avoid damage to business franchise value, which may require voluntary financing transactions such as completion of underwriting pipeline deals and repurchase of securities issued.

6. Businessreduction– The liquidity stress test should incorporate a set of realistic assumptions concerning the institution’s ability to reduce liquidity-draining business activities such as new loan origination. These assumptions should be developed through discussions with business unit management, who will have a view into the level of reduced funding activity that can occur without causing significant reputational problems.

Outputs of The Model

• The outcome of liquidity stress testing, along with the other components of the institution’s liquidity risk measurement framework, provide the foundation for assessing tactical and structural liquidity relative to internally established limits and regulatory expectations. In particular, the liquidity stress test forms an integral part of an institution’s liquidity risk escalation process. The bank’s liquidity limit structure, and in particular the contingency funding plan, should be tied directly to the results of the liquidity stress test. These linkages may exist through, for example, survival horizon metric minimums, minimum available liquidity limits, and stressed liquidity metric limits.

• The liquidity stress test should enable the production of a regular reporting package that contains the following for each of the entities being tested:

1. Stress testing assumptions– Key assumptions include

• overall stress level represented by the scenario
• indication of whether the scenario is systemic, idiosyncratic, or both systemic and idiosyncratic
• documentation of the overall macroeconomic, market, and company-specific events leading to the stress scenario.

2. Liquidity position metrics. The principal measurement outcome of the liquidity stress test is the level of available liquidity relative to net cash outflows under each scenario.

• The exact form of this metric varies across institutions but may be expressed as a percentage of net outflows or as a dollar value relative to a policy minimum.
• Some institutions distinguish between tactical and structural liquidity in measuring the results of the liquidity stress test. The Basel III LCR, for example, provides a thirty day view into available liquidity under stress. To measure the longer-term or structural liquidity position, the institution may calculate a survival horizon relative to a limit. For example, the bank’s policy may be to maintain available liquidity to support twelve months (tactical) of net outflow under a specific stress scenario.

3. Prospective liquidity position metrics – Apart from the current liquidity position, the bank should measure the prospective liquidity profile of the bank over the stress horizon.

• Key indicators of liquidity risk include prospective available liquidity, ratios indicative of wholesale funding dependence (e.g., net non-core funding dependence), and metrics indicative of potential overconcentration in specific funding channels (e.g., percentage of funding from brokered deposits).
• When monitoring prospective liquidity, it is important to highlight any specific stress points along the horizon where survival would require potentially problematic debt issuances, intercompany funding transactions, or capital actions.
• The institution may choose to establish limits for prospective liquidity in addition to current liquidity, for example, maintaining a certain survival horizon throughout the stress test.

4. Capital and performance metrics – It is also important to measure the balance sheet more holistically. For example, assessing the economic impact of the investment portfolio by measuring yield net of a regulatory or economic capital charge enables the institution to assess the tradeoff between low-yielding, low-haircut instruments and higher performance, less liquid ones. Monitoring key capital metrics for each entity ensures that the model captures the impact of any capital actions required to support liquidity during stress period.

• At a minimum, the liquidity stress test should be performed quarterly in order to support review by the asset liability management committee. More advanced banking organizations have made significant investments in building the ability to perform liquidity stress tests more frequently, in some cases even daily.

Governance and Controls

• The specific roles for effective oversight of liquidity stress testing should consist of the following:

• I. Asset-liability committee (ALCO) – The ALCO typically has overall responsibility for the liquidity stress testing framework. Specifically, the ALCO should be responsible for the following:
  • Ensuring the establishment, review, and approval of a liquidity stress testing policy. The liquidity stress testing policy should detail the scenarios to be run, key assumptions, roles and responsibilities, reporting requirements, and limits.
  • Suggesting and approving liquidity risk scenarios, including major changes to liquidity scenarios and/or assumptions
  • Setting liquidity risk policy limits dependent on stress test outcomes and escalating exceptions. For certain limit tiers, escalation may be required to the board of directors.

• II). Treasury. The treasury unit, as the first line of defense, typically has ownership of the liquidity stress test modeling process. Treasury should be responsible for the following:
  • Maintenance of liquidity stress testing procedures.
  • Reviewing and monitoring liquidity characteristics of the institution’s assets and liabilities and making recommendations to ALCO concerning stress testing assumptions.
  • Working with other functions within the organization, in particular business line management, in developing assumptions for customer assets and liabilities. A formal requirement should be established in the liquidity stress testing policy that management reviews the key analytical assumptions of the liquidity stress test at least quarterly.
  • Producing stress test-based liquidity risk reporting.
  • Reconciling liquidity stress test baseline balance sheet data with reports prepared by a group independent of treasury, such as financial control, independent risk management, or middle office.
  • Within large, complex banking organizations, it is expected that multiple treasury units will perform liquidity stress tests for their respective entities. In such cases the corporate treasury group should ensure that the global liquidity stress testing policy establishes a consistent framework of scenarios, assumptions, and model design across the enterprise.

• III) Risk management– The independent risk management function, as the second line of defense, is responsible for providing independent oversight of liquidity stress testing along with the other components of the liquidity risk management program. Specifically, risk management is responsible for the following:
• Administering the liquidity risk stress testing policy
• Reviewing and providing effective challenge of the scenario design and assumptions
• Ensuring the institution’s approach to liquidity stress testing is in line with acceptable industry practices and regulatory rules and guidance
• Reviewing and approving the liquidity stress test-based Limits
• Monitoring of liquidity stress test-based limits
• Ensuring the institution’s ALCO, executive management, and board are kept well- informed of the bank’s liquidity risk profile as indicated by the stress testing results.
• Where multiple risk units within the institution are overseeing liquidity stress testing, the independent risk management function should be responsible for coordinating globally with regional and business unit risk management teams to ensure enterprise-wide consistency.

• IV) Internal audit – Internal audit, as the third line of defense, should periodically review the liquidity stress testing framework, procedures, and controls to ensure compliance with policy, regulatory, and control requirements.
• V) Model risk management – Model risk management is responsible for providing independent validation and changing management governance of the liquidity stress testing model in line with the institution’s model risk management policy. Practices vary among institutions in defining and evaluating models for oversight. It is assumed, however, that the liquidity stress testing model will be assessed as highly critical given its foundational role in monitoring the bank’s risk profile

Liquidity Optimization

• The primary goal of the liquidity stress test is to determine the appropriate size of the liquidity buffer. However, liquidity stress test should also be referenced in developing the composition of the buffer, with the objective of maximizing the efficiency of the liquidity portfolio.

• Liquidity versus yield – Maximizing the yield and/or duration of the portfolio (even under the usual strict investment policy constraints) is likely to be suboptimal for the institution’s return on asset performance as a whole. Conversely, maximizing the liquidity profile of the portfolio at the expense of yield may be equally inefficient. Armed with a robust liquidity stress testing model, treasury should maintain a balance.

• Liquidity versus capital – Depending upon the institution’s economic and regulatory capital framework, a similar trade-off exists for incorporating the capital impact of various portfolio alternatives. Maximizing the investment allocation of some instruments may be suboptimal given the additional regulatory and potentially economic capital requirements associated with these instruments If any additional asset amount gives rise to an additional equity capital requirement under leverage ratio limits, the additional haircut required for higher-risk instruments will also be problematic.

Funding Optimizition

• An important insight provided by the liquidity stress test model is the impact of varying funding sources with differing liquidity characteristics.
• A key objective of the focus on enhancing liquidity stress testing since the financial crisis has been to create an incentive for financial institutions to favor “sticky” funding sources such as retail branch deposits at the expense of “hot” money sourced from wholesale channels. By explicitly modeling the liquidity impact of these funding alternatives, treasury can and should develop a target funding profile that balances liquidity and cost. For example, the superior liquidity profile of commercial deposits linked to treasury management services should serve to bolster the business case for investing in target industry segments with more intensive working capital requirements. Building this linkage requires a funds transfer pricing (FTP) framework that accurately incorporates the stressed liquidity profile of various business segments across the enterprise.

Establishing a Sustainable Infrastructure

• The strongest liquidity stress testing analytical framework will have little value without a data management infrastructure to support it. In order to support efficient and controlled ongoing stress testing and reporting, the institution should maintain an information technology infrastructure that performs automated data collection, aggregation, capturing of market data, report generation, and analytics.

• The challenge for many large, complex financial institutions has been that developing such an infrastructure for liquidity stress data has required significant modifications to existing data warehouse capabilities built largely on general ledger and transactional customer data.

• In building an infrastructure that supports liquidity stress testing, the institution should ensure that several critical requirements are met. These include:

a) Position data collection and aggregation

b) Regulatory report generation

c) Analytics

d) Liquidity dashboard

Integration with Other Risk Models

• Liquidity stress testing should not be performed in isolation without consideration of other related risk frameworks, such as asset liability management (for interest rate risk), capital stress testing, and recovery and resolution planning. These models may employ related assumptions concerning the balance sheet behavior of certain accounts, and developing these assumptions independently is likely to lead to an inconsistent overall risk management framework.

• In theory, the institution should maintain a holistic risk model that assesses the impact on liquidity, capital, and balance sheet structure under a common set of scenarios. In practice, such an approach can be problematic due not only to the modeling complexity involved, but also as a result of the need to develop unique stress scenarios for each risk type.

• The bank should carefully consider the interdependencies and connection points between the liquidity stress model and other risk models.

• Liquidity stress testing and capital stress testing– The liquidity stress test should incorporate any required capital infusions of subsidiary entities.
• For each liquidity stress test scenario, capital impact assumptions must be developed based on the overall market and idiosyncratic conditions assumed to occur under the scenario.
• The capital stress testing framework should include a liquidity stress evaluation to assess the impact of any required liquidity impacts on capital adequacy.
• The institution should perform a liquidity impact analysis to determine whether additional capital impacts may occur through investment portfolio and required funding actions that would cause further deterioration in capital adequacy.

• Liquidity stress testing and asset liability management – Typically, a liquidity impact analysis is not run concurrently with interest rate risk stress testing. This is because an interest rate stress event could have a significant impact on capital but is less likely to have a direct impact on the bank’s short-term liquidity profile.

• Still, a consistent framework should be applied to both interest rate and liquidity stress testing models. For example, if liquidity stress test model assumes that some operational deposits do not run off in a stress event, the interest rate risk model should segment these deposits and assume that their duration would beat least as long as non-operational deposits.
• In particular, the liquidity stress test scenario framework, liquidity risk dashboards, and liquidity risk early warning indicators should not neglect to include the possibility of an interest rate shock and the potential impact such an event would have on indeterminate liabilities. For example, in an environment where rates are historically low and there is significant risk of a yield curve steepening (i.e., the current environment), the institution must carefully consider the impact of deposit disintermediation due to interest rate increases.
• Liquidity stress testing and funds transfer pricing – The FTP framework, while not a risk model, is a strategically important tool for driving business decision making.
• One of the key objectives of any FTP framework is proper pricing of liquidity. The FTP framework should leverage and be consistent with the contingent liquidity requirement for assets and liabilities measured by the liquidity stress test model. For example, if it is determined that a 25% cash buffer is required to support a wholesale operational deposit (to borrow a Basel III LCR assumption), the cost of carrying this buffer should be passed through within the FTP framework.

Conclusion

• The liquidity stress test is a core component of the bank’s liquidity risk framework, and following the financial crisis has become an increased area of scrutiny and expectation among regulators and other stakeholders. While nearly all financial institutions of significant size have a basic liquidity stress testing process in place, there are typically a number of areas of potential improvement. As banks continue to refine and improve their liquidity stress capabilities, they should focus on four areas:

• Ensuring the appropriate scope and structure of the liquidity stress test
• Building the model on robust assumptions
• Improving integration with related risk and performance models
• Automating the process