Solvency Assessment and Management: Stress Testing Task Group Discussion Document 96 (v 3) General Stress Testing Guidance for Insurance Companies 1 INTRODUCTION AND PURPOSE The business of insurance is based on dealing with uncertainty. Therefore, an insurer needs to consider a wide range of possible outcomes that may affect its current and expected future financial position. Although insurers are in many jurisdictions required to conduct stress tests 1, little guidance is provided on how insurers may go about conducting such stress tests. Stress testing is an important management tool to be used in setting business strategy, risk management and capital management decisions. This document sets out general guidance on stress testing for insurance companies. The guidance provided in this document should be treated as a general guide in developing stress tests and should not be seen as prescriptive. 2 STRESS TESTING DEFINED Stress testing is a risk management technique used to evaluate the potentially adverse effects on an institution s current and future financial condition, of a set of specified changes in risk factors, corresponding to unexpected but plausible events. Stress testing includes scenario testing and sensitivity testing 2.1 Scenario testing Scenario testing uses a hypothetical future or relevant historical state of the world to define changes in risk factors affecting an insurer. This will normally involve changes in a number of risk factors, as well as the ripple effects and other impacts that follow logically from these changes and related management actions. Scenario testing is typically conducted over the time horizon appropriate for the business and risks being tested. 2.2 Sensitivity testing Sensitivity testing typically involves a change in a single risk factor (or a limited number of risk factors). It is typically conducted over a shorter time horizon, for 1 Mention here the relevant ICP s and ORSA text
example an instantaneous shock. Sensitivity testing requires fewer resources than scenario testing and can be used as a simpler technique for assessing the impact of a change in risks when a quick response or when more frequent results is needed. The magnitude of the change considered is often relatively small, and is often used to estimate the impact of a change in assumption or the effect of parameter errors in a model. More severe magnitudes may also be considered to evaluate the impact of a large change in a single risk factor on the insurer. Stress testing should be seen not only as complementary to quantitative capital models, but also as an essential tool in the broader risk management system and Own Risk and Solvency Assessment (ORSA) framework. Most quantitative capital models require assigning probabilities to particular outcomes based on past experience. Stress testing in contrast does not require probabilities. It is sufficient to specify the events or sequence of events sufficiently detailed such that the impact on the insurer can be understood and quantified. They are therefore useful as a complement to quantitative models to capture risks that are highly uncertain. Stress tests are also intuitive and useful to communicate with senior managers, board of directors and other stakeholders. Stress tests are explanatory as they do not only give a loss number or risk measure, but clearly define the level of financial stress implied by various events. They help foster an appropriate risk culture since they expose decision makers to potentially inconvenient truths. Insurers who adequately utilise stress tests will thus give due consideration to adverse scenarios to all aspects of business strategy and policy setting. Although much of what follows below will refer to scenario testing, sensitivity tests should be equally considered. 3 PURPOSE OF STRESS TESTING As indicated above, stress testing should be embedded in an insurer s overall risk management system. A risk management system is defined under SAM as the totality of strategies, policies and processes employed to identify, assess, monitor, manage and report on the material risks faced by an insurer 2. A stress testing framework as a whole should be actionable, playing an important role in facilitating the development of risk mitigation or contingency plans across a range of stressed conditions. It should feed into the insurer s decision making process, including setting the insurer s risk appetite, setting exposure limits, and evaluating strategic choices in longer term business planning. 3.1 Identifying risks Stress testing should be included in an insurer s risk management activities at various levels, for example, ranging from informing risk mitigation policies at a detailed or portfolio level to adjusting the insurer s business strategy. In particular, it should be used to address institution-wide risks, and consider the concentrations and interactions between risks in stress environments that might otherwise be overlooked. 2 Discussion Document 71 v5
This also includes risks that currently might not be considered in the quantitative capital models (be it within an internal model or the standard regulatory capital formula). As stress testing allows for the simulation of shocks which have not previously occurred, it can be used to assess the robustness of models to possible changes in the economic and financial environment. As such it can have a role to play in identifying model risk. Stress tests should help to detect vulnerabilities such as unidentified risk concentrations or potential interactions between types of risk that could threaten the viability of the institution, but may be concealed when relying purely on statistical risk management tools based on historical data. Emerging risks can also be identified, defined and understood through the process of stress testing and scenario analysis. Through brainstorming of future potential but as-yet unidentified scenarios as well as reverse stress-testing, emerging risks may be identified. These can then be assessed and monitored along with the known risk types as part of the overall risk management system. 3.2 Assessing risks Stress testing can be particularly useful to quantify risks. Risks may have insufficient data to build statistical models to quantify. Also, risks may not warrant complex quantitative methods from a nature, scale and complexity perspective. In such cases utilising robust stress tests may be more appropriate. Stress tests should complement risk quantification methodologies that are based on complex, quantitative models using backward looking data and estimated statistical relationships. In particular, stress testing outcomes for a particular portfolio can provide insights about the validity of statistical models. Stress testing can also be used to assess the impacts of customer behaviour arising from options embedded in certain products. 3.3 Monitoring risks Regular stress testing using pre-defined stress scenarios can be a useful tool to monitor key risk exposures. The impact of stress scenarios will evolve over time as a result of the changing environment, Regular assessment of previously-defined scenarios can be utilised to gain a deepened understanding of how the impact of such exposures has evolved over time. This includes emerging risks discussed in section 3.1. 3.4 Managing risks Once a scenario or stress has been assessed, consideration should be given to any actions or controls which could be put into place now to mitigate the impact of such a scenario. Due consideration should be given to the cost of such controls, and the cost considered in light of the insurer s risk appetite. An integral part of stress testing is the consideration of management actions taken in times of stress. When considering the impact of scenario, it may be possible to reduce the impact by certain decisions or actions. The typical example of such action is the reduction or elimination of non-vested bonuses on with-profit life products. Insurers should however be wary of what management actions are possible in times of stress, and any constraints on such actions. Examples of constraints may include legislative (e.g. policyholder protection rules) and economic
(unavailability of additional capital in the market, or terms unfavourable). [add knock on effects on reputation of actions] When such actions assume decision-making by senior management or the board, such decisions should be pre-agreed and formalised as much as possible, for example by writing such actions into policy. The ORSA requires that insurers consider and quantify risks more broadly than those considered in the regulatory capital requirement. One such risk which is not fully catered for in the standard formula is liquidity risk. Stress testing should be a central tool in identifying, measuring and managing liquidity risks. In particular, the insurer s liquidity profile and the adequacy of liquidity buffers should be assessed in both insurer-specific and market-wide stress events. An insurer s ORSA furthermore requires interaction between its risk management system and its capital management activities. Stress testing should form an integral part of insurers internal capital management where rigorous, forward-looking stress testing can identify severe events, including a series of compounding events, or changes in market conditions that could adversely impact the insurer. This should influence the projections of Own Funds made in the ORSA. 4 DEVELOPING STRESS TESTS An insurer should regularly maintain and update its stress testing framework. The effectiveness of the framework, as well as the robustness of individual components, should be assessed regularly and independently. The stress testing framework should deal with types of stress tests and identification of scenarios, assessment of the impact of such stresses and scenarios, mitigating actions assumed in the assessment, and monitoring and reporting of stresses and scenarios. 4.1 Identification of scenarios Stress testing programs should take account of views from across the organisation and should cover a range of perspectives and techniques. Defining appropriate events is as intellectually challenging as developing a quantitative capital model and requires appropriate knowledge and experience. It should employ specialists from fields relevant to the insurer. This includes, but is not limited to, the following persons or functions: the risk function, the actuarial function, senior management, the Board, reinsurers and rating agencies. Academics or specialists from related industry bodies or organisations may also be called upon. Scenarios selection Stress tests should cover a range of risks and business areas, as well as for the business as a whole. An insurer should be able to integrate effectively across the range of its stress testing activities to deliver a complete picture of the risks on the business as a whole. Comprehensive stress testing programs should consider the insurer s most material risks and aim to take into account system-wide interactions and feedback effects (e.g. second order and macroeconomic effects). This means that the identified scenarios should not only consider the initial strain, but also the sequence of effects that can play out over months and years. However, it is
difficult, if not impossible, to think through all the possible ramifications of such an event. The purpose of assessing a scenario is less about predicting a future event than it is about getting an insurer to think in advance about certain types of events and to be prepared if a similar (but not necessarily identical) scenario was to occur. It is therefore sufficient that the formulation of the scenario captures the essence of the effects of an initial disruptive event. Where relevant and material, such risks may include: Credit risk (i.e. counterparty and reinsurance risk) Market risk Insurance risk (e.g. mortality, morbidity, claim frequency and severity, persistency and lapse risk) Liquidity risk Operational and legal risk Concentration risk Contagion risk Risk to reputation Securitisation risk New business risk Regulatory risk Inflation risk Scenario selection should be conducted flexibly and imaginatively in order to improve the likelihood of identifying hidden vulnerabilities. A failure of imagination could lead to an underestimation of the likelihood and severity of extreme events and to a false sense of security about an insurer s resilience. The institution should assess the impact of severe shocks and periods of severe and sustained downturns, including its ability to react over the appropriate time horizon for the business and risks being tested. Further, stress tests should identify and respond to potential changes in market conditions that could adversely impact an institution s exposure to risk concentrations. Reverse stress testing A stress testing program should also determine what scenarios could challenge the viability of the insurer (reverse stress tests). Such tests may be useful in uncovering hidden risks and interactions among risks. The primary use of reverse stress tests is as a risk management tool to improve business planning and risk management rather than to inform decisions on appropriate levels of capital. However, reverse stress testing may result indirectly in changes to the levels of capital held by firms. For example, if a firm s reverse stress testing identifies business model vulnerabilities that have not previously been considered, but which the firm expects to survive according to its risk appetite/tolerance, the firm may decide to take mitigating action or hold a different amount of capital. A reverse stress test starts with a specified outcome that challenges the viability of the insurer. One example of such an outcome would be that over a short time period, the insurer incurs a very large loss that challenges its viability. The analysis would then work backwards (reverse engineered) to identify a scenario or combination of scenarios that could bring about such a specified outcome. The reverse stress test
induces insurers to consider scenarios beyond normal business settings that would include events with contagion and systemic implications. As described earlier in the document, a stress testing program should also determine what scenarios could challenge the viability of the insurer (reverse stress tests). A firm s business model is described as being unviable at the point when crystallising risks causes the market to lose confidence in the firm. A consequence of this would be that counterparties and other stakeholders would be unwilling to transact with or provide capital to the firm and, where relevant, that existing counterparties may seek to terminate their contracts. Such a point could be reached well before a firm s regulatory capital is exhausted. Specific consideration should be given to important inter-relations between various risk factors. For a long-term insurer, changes in economic conditions can significantly affect policyholder behaviour such as lapse rates, utilisation of options within an insurance contract, and morbidity and recovery rates. For a short-term insurer, changing economic conditions will not only influence investment income and company expenses, but can also lead to higher claims and loss reserves, particularly in times of inflation. The inter-relations of various factors will depend upon the insurer s products, its investment policy and its approach to managing its business. A critical goal for insurers is to identify situations in which the assumed normal pattern of inter-relationships breaks down due to a change in the business environment. Historical scenarios A historical scenario is based on an observed historical event. A major advantage of a scenario based on a historical event is that it can be more easily communicated since the event has actually occurred. As a further advantage, since such a scenario is based on an observed event, data with regards to its short-, medium-, and long-term impact might be available. In particular, the impact of the event on other risk factors, such as interest rates, equity markets, and inflation can be studied, as can its pre-history, i.e. the history leading up to the event. Clearly, the circumstances surrounding the historical event will be different to the current situation. Therefore, in constructing the scenario, appropriate adjustments need to be applied. In many scenarios, financial values will have to be adjusted for inflation to make the values more consistent with current values. Great care has to be taken when a historical scenario relates to the financial markets. Financial markets are constantly evolving and an event that had little impact in the past might have a big impact now or vice versa. Other examples of factors that require adjustment in historical scenarios include: Changes in population movement Medical advances New technologies Globalised and more closely linked financial markets New asset classes owned by insurers Management behaviour and incentive schemes Different valuation methods and bases used, making historical data not directly comparable to current values
Such adjustments cannot realistically consider all differences between the present situation, and that of the historical scenario. The idea is to use historical events as best as possible, given the limitations, to assist in dictating scenarios. Synthetic scenarios In contrast to historical scenarios, synthetic scenarios describe events which have not yet been observed and which thus can be more easily tailored to a specific situation of the insurer. Synthetic scenarios require more assumptions than historical scenarios. For this reason they are subject to more questioning and can be more difficult to communicate and discuss both within the company and with third parties. Formulation of scenarios A scenario needs to be formulated such that the financial impact on the insurer can be calculated. The first step in formulating a scenario is to explain it in a concise, understandable narrative that outlines the initial event and the potential effects cascading from it. The narrative should explain the key assumptions and simplification that have been made. It should focus on the qualitative aspects of the scenario but also include the major quantitative assumptions. The qualitative formulation describes the effects of the initial event over the duration of the scenario. If relevant, this also includes a description of the effects not only of the risk factors, which affect the firm directly (e.g. mortality rates, claims, interest rates, etc.), but also the effects on other market participants. The sequence in which events occur can be of major consequence for the financial impact on the insurer. The qualitative description therefore should not only contain the sequence of effects but also a time-line. This is particularly relevant for scenarios, which evolve over months and years or in which the effects of actions that management would take are to be included in the evaluation of the scenario. Many risk factors which enter into scenarios exhibit dependencies or relationships. When formulating a stress scenario, it is important to incorporate dependencies as they exist in stressful circumstances. Experience has shown that in many situations dependencies under stressed situations are different from dependencies under normal situations. In practice, dependencies under stress will usually not have been observed. It would therefore then be necessary to model them based on informed judgement. It should also be noted that the Solvency Capital Requirement (SCR) calculation under the standard formula of the Solvency Assessment and Management (SAM) regime is a defined stressed scenario. Risks not covered in the regulatory capital formula Certain risks an insurance company faces cannot be appropriately quantified, since their likelihood of occurring is highly uncertain or there is no underlying scientific theory. However, they do have quantifiable consequences that can have strong effects on a firm s financial condition. An example of a risk with a highly uncertain probability of occurrence is reputation risk. The likelihood of financial loss due to loss of reputation depends on the firm s strategy, the behaviour of its senior management, operational risks etc. However, while it is difficult, if not impossible, to give a reasonable close estimate of the probability of this occurring, the financial damage once such a reputational loss
happens can be expressed numerically. Such damage may affect policyholder behaviour (lapse rates), new business volume, employee turnover or productivity, relations with suppliers, and regulatory actions. All of these effects will have financial consequences. In order to consider the effects of such a risk, one must construct a scenario that describes the situation and its impact on various financial elements of the insurer s operations. 4.2 Evaluation of scenarios The evaluation of a given scenario consists of the quantification of its impact on the insurer s balance sheet. The quantification of the impact depends on the valuation framework. In many cases, it is important to gain insight into the insurer s financial state based on different valuation frameworks, e.g. economic valuation, statutory valuation. If the scenario continues over several years, the evaluation should include the quantification of the effects of the scenario over the entire duration of the event. The main purpose of a scenario framework is not merely to obtain quantitative results, but to gain insight such that the insurer can manage its risk in such a way as to cope with unexpected events. The insurer should not merely quantify the effects of the scenario but should also formulate contingency plans. Impact measure The impact of stress tests is usually evaluated using one or more measures. The particular measures used will depend on the specific purpose of the stress test, the risks and portfolios being analysed and the particular issue under examination. A range of measures may need to be considered to convey an adequate impression of the impact. Typical measures used are: Asset and liability values Level of impaired assets and write-offs Profit and loss Required and available regulatory capital Economic capital Liquidity and funding gaps 4.3 Allowance for management action and other mitigations Stress testing should facilitate the development of risk mitigation or contingency plans across a range of stressed conditions. The performance of risk mitigating techniques, like reinsurance, hedging, netting and the use of collateral, should be challenged and assessed systematically under stressed conditions when markets may not be fully functioning and multiple institutions simultaneously could be pursuing similar risk mitigating strategies. Stress testing should also reflect constraints on management action and should not place undue reliance on the timeliness of mitigating actions. Factors such as company history and philosophy, competitive position and policyholders reasonable expectations can have important effects on the timing of proposed action. 5 STRESS TESTING GOVERNANCE A strong governance structure is required given the importance of the stress testing. Since the stress test development and maintenance processes often imply
judgemental and expert decisions (e.g. assumptions to be tested, calibration of the stress, etc.), the independent control functions such as risk management and internal audit should also play a key role in the process. In particular, there should be an independent review (e.g. by internal audit) of the adequacy of the design and effectiveness of the operations of an institution s stress testing programs. The Board and senior management are responsible for and should ask probing questions on stress testing. This challenging is a vital part in ensuring the business derives value from this process and will require Board members and senior management to become knowledgeable in this area. Assessments of effectiveness should be qualitative as well as quantitative, given the importance of judgements and the severity of shocks considered. Areas for assessment should include the effectiveness of the program in meeting its intended purposes, documentation, development work, system implementation, management oversight, data quality and hypotheses and assumptions used. The Board and Senior Management are responsible for ensuring that the stress testing is administered by personnel with the relevant skills and expertise to do so. The Task Group recognises the potential need for a document (internal to the insurer) that describes the principles underpinning the process whereby stress testing is performed. Thus insurers and insurance groups should develop a policy specifically relating to stress testing. This policy should form part of the overall policy suite and governance framework. The stress testing and documentation around stress testing should be performed in such a way that it can be reviewed with reasonable effort by a third party. As an example, internal audit may evaluate it from a process perspective. Thus the documentation should be complete, accurate and present clear audit trails. Documentation should not be overly and unnecessarily complex. Insurers should have written policies and procedures governing the stress testing program. The operation of the program should be appropriately documented. The assumptions and fundamental elements for each stress testing exercise should also be appropriately documented, including the reasoning and judgements underlying the scenarios chosen and the sensitivity of stress testing results to the range and severity of the scenarios. The level of documentation should be based on the nature and purposes of the stress testing.
SELECTED REFERENCES The following list includes some of the works referred to in the paper. Financial Services Authority (FSA). Policy Statement 09/20 Stress and Scenario Testing (December 2009). International Association of Actuaries (IAA). Paper on Stress Testing and Scenario Analysis. International Association of Insurance Supervisors (IAIS). Stress Testing by Insurers Guidance Paper (October 2003). Office of Superintendent of Financial Institutions Canada (OSFI). Guideline Stress Testing (December 2009).