Month: February 2010

Dynamic financial analysis (DFA) is an application of mathematical modelling to businesses. DFA models the key elements that impact an organisation’s operations and simulates thousands of potential situations, determining the firm’s financial condition for each outcome. The output from DFA is the distribution of potential financial results for the next few years.

DFA is used for a variety of reasons, including solvency regulation, assigning financial ratings, evaluating a change in operations, and determining the value of an organisation for an acquisition. All DFA models are based on a number of key assumptions that govern how the values of future financial variables are determined. Key assumptions include interest rates, inflation, stock returns and catastrophe claims. DFA models generally have several interrelated modules that interact to generate the results. Separate modules for an insurer can include underwriting gains and losses, catastrophe losses, reinsurance agreements, investments, and taxation. When all the modules are combined, the total after tax operating statement and balance sheet for the insurer are calculated for the next few years. A single iteration is likely to require thousands of random variables and tens of thousands of calculations, with each run of a DFA programme performing thousands of iterations to determine a probability distribution for future financial positions. This type of simulation could not be done prior to the development of powerful computers that perform these operations.

Output from a DFA model is used to determine a firm’s financial condition based on the likelihood of its developing financial difficulties on the basis of its current operating conditions. For each iteration associated with financial difficulties, all the relevant variables can be captured and analysed. For example, an insurer could discover that most of the incidents of financial impairment are associated with large catastrophe losses. This could lead the insurer to revise its reinsurance contracts or change the geographical areas in which it operates.

Despite the power of an effective DFA model, care needs to be used in applying these models. DFA, as any model, is a simplified representation of reality. Many factors that can influence results are not included in the models. Thus, all the uncertainty is not reflected in the model; actual results will vary more widely than the model will indicate. Models are built on the basis of what has happened in the past, not on what new conditions can arise. Firms that do not recognise this inherent uncertainty, and accept levels of leverage that are based on the results of a DFA model, are more exposed to risk than they and their regulators realise. Also, situations can change so that factors not included in the model will need to be added. DFA is dynamic in the sense that it will continually evolve. In general, DFA models can be useful as a guide for firms, but they must be used with a healthy dose of skepticism.

This article is an edited version of
an entry in the “Encyclopedia of Quantitative Risk Analysis and
Assessment”, Copyright © 2008 John Wiley & Sons Ltd. Used by
permission.

www.wiley.com

Reliance on models to price, trade, and manage risks carries risk. Models are susceptible to errors.

In liquid and more or less efficient securities markets, the market price is, on average, the best indicator of the asset’s value. In the absence of liquid markets and price discovery mechanisms, there is no alternative than theoretical valuation models to mark-to-model the position, to assess the risk exposure along the various risk factors, and to derive the appropriate hedging strategy.

The pace of model development has accelerated to support the rapid growth of financial innovations such as caps, floors, and others. These have been made possible because of developments in theory, which allow a better capture of financial risks. At the same time, these models would have never been implemented in practice, had the growth in computing power not accelerated so dramatically. The main causes of model risk are model error and implementing a model wrongly.

Model error
A model is incorrect if there are mistakes in the analytical solution. A model is also incorrect if it is based on wrong assumptions about the underlying asset price process. Finance is littered with examples of trading strategies based on shaky assumptions – some model risks are really just a formalisation of this.

The most frequent error in model building is to assume that the distribution of the underlying asset is stationary, when it changes. The ideal solution would be to acknowledge that volatility is stochastic and to develop an option-pricing model, but option-valuation models become difficult when any sort of stochastic volatility is included.

Implementing a model wrongly
With models that require extensive programming, there is always a chance that a bug may affect output. Some implementations rely on techniques that exhibit inherent errors and limited validity. Many programmes that seem error-free have been tested only under normal conditions.

For models evaluating complex derivatives, data are collected from different sources. The implicit assumption is that for each period, the data for all relevant assets and rates pertain to exactly the same instant, and thus reflect simultaneous prices. Using nonsimultaneous price inputs may be but can lead to wrong pricing.

When implementing a pricing model, tools are used to estimate parameters. But how frequently should input parameters be refreshed? Should the adjustment be made on a periodic basis, or should it be triggered by an event? Should parameters be adjusted according to qualitative judgements, or should these be based on statistics?
Statistical estimators are subject to errors involving inputs. A major problem in the estimation procedure is the treatment of outliers. Are the outliers really outliers, in the sense that they do not reflect the true distribution? Or are they important observations that should not be dismissed? The results of the estimation procedure will be vastly different depending on how observations are treated. Each bank may use a different procedure to estimate parameters. Some may use daily closing prices, while others may use transaction data.

The quality of a model depends on the accuracy of inputs and parameter values. This is particularly true in the case of new markets, where best-practice procedures and controls are evolving. Volatilities and correlations are the hardest parameters to judge.

Most institutions use internal data as well as external databases. The responsibility for accuracy is often not clearly assigned.

Adding observations improves the power of tests and tends to reduce errors; but the longer the sampling perio, the more weight is given to obsolete information.

The gap between the bid and ask prices may be large enough to complicate the process of finding a single value. Choices made about the price data at the time of data selection can have a major impact on the model.

How can we mitigate model risk?
One way is to invest in research to improve models and to develop statistical tools. It is critical for an institution to keep up with developments.

An even more vital way of reducing model risk is to establish a process for independent vetting of how models are constructed.

The role of vetting is to assure management that any model for the valuation of a given security proposed is reasonable. It provides assurance that the model offers a reasonable representation of how the market itself values the instrument, and that the model has been correctly implemented.
1. Documentation
This should be independent of any implementation, such as a spreadsheet or code, and should include:
(a) the term sheet or;
(b) a statement of the model
2. Soundness of model
The vetter needs to verify that the mathematical model is a reasonable representation of the instrument.
3. Independent access to financial rates
The vetter should check that the bank’s middle office has independent access to an independent market-risk-management rates database.
4. Benchmark modelling
The vetter should develop a benchmark model based on the assumptions and on deal specifications. The vetter may use an implementation that is different from that proposed.
5. Health check and stress test
6. Build a formal treatment of risk into the overall risk-management procedures

Large trading profits tend to lead to large bonuses for senior managers, and this creates an incentive for these managers to believe the traders who are reporting the profits. Often, traders use their expertise in formal pricing models to confound internal critics, or they may claim to have some sort of profound insight. Senior managers should approach any model that seems to record or deliver above-market returns with a healthy skepticism, to insist that models are made transparent, and to make sure that all models are independently vetted.

This article is an edited version of an entry in the “Encyclopedia of Quantitative Risk Analysis and Assessment”, Copyright © 2008 John Wiley & Sons Ltd. Used by permission.

www.wiley.com

ALM is universally defined as a comprehensive analysis of the asset portfolio in light of current liabilities and future cash flows of a going-concern company, incorporating existing asset and liability portfolios as well as future premium flows. ALM also considers additional risk factors beyond interest rate changes such as inflation, credit, and market risk. ALM also considers actions beyond the characteristics of a fixed income portfolio and seeks to identify and exploit hedges.

Insurance companies can benefit from a more integrated analysis of the asset and liability portfolios in seeking better risk-return decisions. An enterprise-wide analysis of potential risks and rewards affords an opportunity to analyse the company’s investment portfolio and underwriting portfolio. Since insurance liabilities are far less liquid than assets, such analysis and management activity tend to focus on adjustments to the investment portfolio, given the constraints of the reserves and underwriting portfolio, to improve the risk-return characteristics. In this respect assets can be thought of as a way to hedge liability risk. However, management activity need not be confined to fine-tuning investment strategy. Future underwriting considerations, along with other hedges such as reinsurance, are risk-management variables at their disposal.

Venter et al. presented a series of simple examples illustrating that the optimal risk-return portfolio decisions are very different as the asset and liability considerations become more realistic and complex. The authors started with a standalone asset portfolio, then with adjustments added a constant fixed duration liability, a liability that varied as to time and amount, and then added consideration of cash flows from current underwriting. As the various layers of complexity are added to the illustration, the nature of the inherent risks changes, as does the optimal investment portfolio.

The study did not address tax considerations, which can have a profound impact on investment decisions. Recent studies have found that insurers consider cyclical changes in the portfolio between tax-exempt and taxable fixed income securities over the course of the underwriting cycle to be one of the principle drivers in investment strategy. In addition to the integration of underwriting and investment results, such strategies rely on reallocation of assets to maximise income while avoiding alternative minimum taxes (AMT).

Consideration of equities, too, adds complexity and richness to the asset–liability analysis. Equities are considered risky in their own right and will imply a potentially worse downside risk to capital. Some believe that equities may provide a better inflation hedge for liabilities in an increasing loss cost environment. This proposition may be tested through the enterprise risk model, although the conclusion will be sensitive to input assumptions of the incorporated macroeconomic model.

In 2002, the Casualty Actuarial Society Valuation, Finance, and Investment Committee (VFIC) published a report testing the optimality of duration matching investment strategies for insurance companies. VFIC attempted to tackle Venter’s most complex scenario discussed above.

Where Venter et al., focused on changes in GAAP pretax surplus changes as the risk measure, VFIC looked at several different risk measures on both a statutory and a GAAP basis. Return, too, was considered on both accounting bases. In doing so, VFIC’s conclusion as to optimality was what one might expect in the real world: it depends. Duration matching was among a family of optimal strategies, but the choice of specific investment strategies was dependent on the company’s choice of risk metrics, return metrics, and risk-return tolerances or preferences.

An asset–liability modelling approach
It has been asserted that an enterprise-wide model is the ideal way to model and ultimately manage an insurance company investment portfolio. ALM makes for an excellent application of such an integrated model.
1.Start with models of asset classes, existing liabilities, and current business operations.
2.Define risk metrics for the analysis.
3.Similarly, management must define what constitutes return.
4.Consideration must be given to the time horizon of the analysis and the relevant metrics.
5.The model will have to consider relevant constraints.
6.The model should be run for a variety of investment strategies, underwriting strategies, and reinsurance options under consideration.
7.An efficient frontier – a plot of the return metric versus the risk metric – can be constructed across the various portfolio scenarios.
8.Since liabilities are more illiquid, the asset–liability analysis and management can be largely asset centric given the existing liabilities.
9.Having selected a targeted point on an efficient frontier and a companion reinsurance strategy, simulation output should be analysed to identify those scenarios where even the preferred portfolio(s) performed poorly.

Future research
While enterprise modelling is perhaps the only way to adequately address asset–liability management issues, there are a number of real-world issues that are subject of continuing research. Correlations can materially alter the risk of the optimal portfolio. Also, models of unpaid losses have not been developed as explanatory models. That is, unlike asset models, reserving models do not predict future loss payments with parameters linking losses to economic indices. Inflation sensitivity is often hypothesised on an accident year, a calendar year, or a payment year basis, but rarely explicitly developed from historic economic data and projected on the basis of, say, an economic scenario generator.

This article is an edited version of an entry in the “Encyclopedia of Quantitative Risk Analysis and Assessment”, Copyright © 2008 John Wiley & Sons Ltd. Used by permission.

www.wiley.com

The word actuary derives from the Latin actuarius, who was the business manager of the Senate of Ancient Rome. It was applied to a mathematician of an insurance company in 1775 in the Equitable Life Insurance Society of London.

Actuarial science
This provides a structured and rigid approach to modelling and analysing uncertain outcomes of events that may impose or imply losses or liabilities upon individuals or organisations.

Given that uncertainty is a main characteristic of actuarial events, it follows that probability must be the cornerstone in the structure of actuarial science. Probability in turn rests on pure mathematics.

In order to enable probabilistic modelling of actuarial events to be a realistic and representative description of real-life phenomena, understanding of the “physical nature” of the events under consideration is a basic prerequisite. Pure mathematics and pure probability must therefore be supplemented with and supported by the sciences that deal with such “physical nature” understanding of actuarial events.

Actuarial practice
The main practice areas for actuaries can broadly be divided into: life insurance and pensions; general/nonlife insurance; and financial risk.

There are certain functions in which actuaries have a statutory role. Evaluation of reserves in life and general insurance and in pension funds is an actuarial process, and it is a requirement under the legislation in most countries that this evaluation is undertaken and certified by an appointed actuary. The role of an appointed actuary has long traditions in life insurance and in pension funds. A similar requirement has been introduced by an increasing number of countries since the early 1990s.

The involvement of an actuary can be required as a matter of substance in other functions. Then there are functions where actuarial qualifications are neither a formal nor a substantial requirement, but where actuarial qualifications are perceived to be a necessity.

Life insurance and pensions
Assessing and controlling the risk of life insurance and pension undertakings is the origin of actuarial practice and the actuarial profession. The success in managing this risk comprises the following:
•understanding lifetime as a stochastic phenomena, and modelling it within a probabilistic framework;
•evaluating the diversifying effect of aggregating lifetime of several individuals into one portfolio;
•estimating individual death and survival probabilities.

By analysing historic data and projecting future trends, life insurance actuaries constantly maintain both first-order and second-order bases. Premium and premium reserve valuation tariffs and systems are built on the first-order basis. Over time emerging surplus is evaluated and analysed by source, and in due course reverted as policyhold¬ers’ bonus.

General insurance
Over the years actuaries have attained a growing importance in the running of the nonlife insurance operations. The basis for the insurance industry is to accept economic risks. An insurance contract may give rise to claims. Both the number of claims and their sizes are unknown to the company. Thus insurance involves uncertainty and here is where the actuaries have their prerogative; they are experts in insurance mathematics and statistics.

Finance
Financial risk has grown to become a relatively new area for actuaries. Those who practice here are called “actuaries of the third kind”.

It is different from ordinary insurance risk in that increasing the size of a portfolio does not in itself provide a diversification effect. For actuaries it has been a disappointing that the law of large numbers does not come into assistance.

The most fundamental and innovative result in the theory of financial risk/mathematical finance is probably that risk associated with contingent claims can be eliminated by appropriate portfolio management.

This theory is the cornerstone in a new practice field, called financial engineering. Activities include quantitative modelling and analysis, funds management, interest rate performance measurement, asset allocation, and model-based scenario testing. Actuaries may practice financial engineering in their own right, or they may apply financial engineering as an added dimension to traditional insurance-orientated actuarial work.

A field where traditional actuarial methods and methods relating to financial risk are beautifully aligned is asset liability management (ALM). The overriding objective of ALM is to gain insight into how money is best allocated among given financial assets, in order to fulfill specific obligations represented by a future payment stream. The analysis of the obligation’s payment stream rests on traditional actuarial science, the analysis of the asset allocation problem falls under the umbrella of financial risks, and the blending of the two is a challenge that requires insight into both and the ability to understand and model how financial risk and insurance risk interact. ALM is today a key component in the risk management of insurance providers, pension funds, and other financial institutions.

A new challenge on the horizon is the requirement for insurers to prepare financial reports on a market-based principle, which the International Accounting Standards Board has had under preparation for some time. In order to build and apply models and valuation tools that are consistent with this principle, actuaries will be required to combine traditional actuarial thinking with ideas and methods from economics and finance.

This article is an edited version of an entry in the “Encyclopedia of Quantitative Risk Analysis and Assessment”, Copyright © 2008 John Wiley & Sons Ltd. Used by permission.

www.wiley.com

Each year consumer products are involved in millions of injuries and thousands of fatalities. Responsibility for this rests with the manufacturers. Most developed countries have also established government agencies that provide regulatory oversight. In the US, the Consumer Product Safety Commission (CPSC) maintains regulatory jurisdiction over more than 15,000 types of consumer products.

Consideration for public safety can improve product development and product quality. Formal tools of quantitative risk assessment add value by improving decision making at different stages.

Product characteristics and analysis of historical data on product-related injuries and deaths can inform safety assessment of a product design before it is manufactured.

Unintended declines in the quality of consumer items can pose safety hazards to consumers, causing unexpected financial burdens and potential legal liabilities to manufacturers. At the beginning of the production process, companies may implement acceptance sampling plans to ensure the quality of raw materials or components. The acceptable quality limits (AQLs) used in such plans can be chosen on the basis of a quantitative risk assessment. Standard tools for statistical process control are routinely deployed by manufacturers to provide checks on quality. A motivating principle is that successful business operations require continuous effort to reduce variation in process outputs. End-of-line testing is typically performed before manufacturers approve products for shipping to customers. In some industries such testing may be required by government regulation.

Information about the performance of a product after release to consumers comes from returns, warrantee programmes, and complaints. Records of incidents involving failure of a product during operation typically include information such as model and serial number of the unit, production date or date code, incident date, and location. In monitoring field performance, a manufacturer needs to determine whether an adverse change has occurred and, if so, assess its implications for product reliability and safety. If a background level of event risk is accepted, statistical methods can determine whether adverse events can be explained by random variation or require attention. Production, shipping, and sales records typically provide the raw data for such analyses.

Confirmation of a problem raises a series of questions: How bad is it? Do field data point to a particular production period or facility? A quantitative risk assessment can address these questions using statistical estimation and hypothesis testing. Risk analysts review the information about production history and seek changes in time period, plant, process, equipment, design, or supplier that are associated with subsequent problems. Parametric statistical models are frequently used to fit time-to-failure distributions in engineering applications.

Product recalls require identification of the units affected, typically by time or source of production, and determination of whether affected units will be repaired or replaced. Increasing globalisation challenges manufacturers and regulators of consumer products, as a product may contain components from different nations and be sold all over the world.
The process for implementing corrective action varies from country to country. In 1997 the US CPSC adopted a “Fast Track Product Recall Programme” for reports filed according to Section 15(b) of the Consumer Product Safety Act (CPSA). This programme requires manufacturers, distributors, and retailers of consumer products to notify the commission of certain defects, unreasonable risks, or noncompliance with voluntary or mandatory stan¬dards. Under the CPSC Fast Track programme, the staff refrains from making a preliminary determination when firms report and implement an acceptable corrective action plan. The plan submitted to CPSC must describe the recall action (refund, repair, or replace) that the company will take to eliminate the identified risk, and provide sufficient information on the product design, incident, and testing to allow the CPSC staff to determine whether the proposed action can correct the identified problem.

An effective risk assessment must link measured levels of risk to specific corrective actions. Supporting tools have been developed, particularly in the EU.

A community rapid information exchange system known as RAPEX assesses the risk of potentially hazardous consumer products by first considering (a) the probability of health damage from regular exposure, and (b) the severity of injury from the product. Depending on the probability and severity of damage the RAPEX method then classifies the overall gravity of an adverse outcome on a five-point ordinal scale ranging from very low to very high. The final judgment of whether the risk requires corrective action considers three additional factors: (a) the vulnerability of people exposed to the product, (b) whether the hazard is obvious to nonvulnerable adults, and (c) whether the product has adequate warnings and safeguards. The RAPEX methodology is sometimes viewed as the preferred approach in countries adopting a risk-averse approach to consumer product safety based on such rough scoring.

Although consumer products are required to be safe, safety does not mean zero risk. A safe product is one that has reasonable risks, given the magnitude of the benefit expected and the alternatives available. In managing consumer product risk, quantitative risk assessment and associated statistical methods are used to frame substantive issues in terms of estimable quantities and testable hypotheses, to extract infor¬mation on product performance from data on field incidents and manufacturing records, and to com¬municate findings to upper management, regulatory authorities, and consumers.

This article is an edited version of an entry in the “Encyclopedia of Quantitative Risk Analysis and Assessment”, Copyright © 2008 John Wiley & Sons Ltd. Used by permission.

www.wiley.com

Enterprise risk management (ERM) is a recent technique, practiced increasingly by large corporations in industries throughout the world. Sensible risk management flows from the recognition that a dollar spent on managing risk is a dollar cost to the firm, regardless of whether this risk arises in the finance arena or in the context of a physical calamity such as fire. ERM thus proposes that the firm addresses these risks in a unified manner, consistent with its strategic objectives and risk appetite.

Most corporations adopt the definition of ERM proposed by the Committee of Sponsoring Organisations of the Treadway Commission (COSO) in their 2004 ERM framework. It intended to establish key concepts and techniques for ERM. In this framework, ERM is defined as “a process, affected by an entity’s board of directors, management and other personnel, applied in strategy setting and across the enterprise, designed to identify potential events that may affect the entity, and manage risk to be within its risk appetite, to provide reasonable assurance regarding the achievement of entity objectives”. This definition highlights that ERM reaches the highest level of the organisational structure and is directed by corporations’ business strategies. The concept of risk appetite is crucial. Risk appetite reflects a firm’s willingness and ability to take on risks in order to achieve its objectives.

As a rising management discipline, ERM varies across industries and corporations. The insurance industry, financial institutions, and the energy industry are among the industry sectors where ERM has seen relatively advanced development in a broad range of corporations. Recently, even the public sector, are becoming aware of the potential value of ERM and risk managers are increasingly bringing it to top executives’ agendas.

Notwithstanding the attractiveness of ERM conceptually, corporations are often challenged to put it into effect. One of the main challenges is to manage the totality of corporation risks as a portfolio in the operational decision process, rather than as individual silos, as is traditionally done.

Operationalisation of ERM
The core of the challenge lies in operationalising ERM. Integration of risks is not merely a procedure of stacking all risks together, but rather a procedure of fully recognising the interrelations among risks and prioritizing risks to create true economic value. Important components of this procedure include risk identification, risk measurement, risk aggregation/other modelling approaches, risk prioritisation, and risk communication.

The four major categories of risks considered under an ERM framework are hazard risk, financial risk, operational risk, and strategic risk.

Under ERM, the identification of individual risks in different categories should facilitate successive prioritisation and integration of risks to best achieve business objectives within the corporation’s risk appetite. Any event that may adversely affect the corporation’s achievement of its objectives is considered a risk under ERM. Proper objective identification is a prerequisite for risk identification. For example, business objectives can be described by certain key performance indicators (KPIs), which are usually financial measures such as ROE, operating income, earnings per share (EPS), and other metrics for specific industries, eg, risk adjusted return on capital (RAROC) and risk-based capital (RBC) for financial and insurance industries. Risks are then recognised by means of these company performance metrics.

Prioritisation
To realise effective risk integration, ERM also promotes risk prioritisation. Risk prioritisation stems from the fact that risks are not equally important to corporations. Prioritisation should reflect different aspects of the company’s strategies and risk-management philosophy, eg, cost to tolerate that risk, reduce it, elicit and apply management’s risk preferences, etc.

ERM and compliance
ERM at first arises from corporations’ efforts to comply with laws and regulations. To this end, it is seen more as an efficient internal control process. Within a corporation, it is often conducted with internal control functions and supervised by internal auditors. The most significant regulatory forces responsible for the rise of ERM are the Sarbanes Oxley Act of 2002, the Basel Capital Accord II, and rating criteria set forth by rating agencies such as Standard & Poor’s (S&P).

ERM future – value creation
ERM practices may have been initially driven by compliance needs, but developments should continue to serve as an internal control function for better corporate governance. One common objective for corporations is to maximise firm value. ERM provides a framework for corporations to consciously optimise risk/return relationships. This optimisation is achieved through the alignment of corporate strategic goals and risk appetite. At the operational level, the alignment guides virtually all activities conducted by the corporation. Specific risks are identified and measured. They are prioritised and integrated by recognising the interrelations and relative influences affecting different risky outcomes. Risk management strategies are developed for the entire portfolio of risks and their effects are assessed and communicated.

This article is an edited version of an entry in the “Encyclopedia of Quantitative Risk Analysis and Assessment”, Copyright © 2008 John Wiley & Sons Ltd. Used by permission.

www.wiley.com