Economic Capital Modelling Flashcards
What are the 5 types of risk assessment and quantification tools?
1) risk assessments
2) loss event database
3) KRI
4) risk analytical models
5) economic capital models
Lam - ERM Textbook - pg. 369
Define required economic capital
The capital required to support a business with a certain probability of defualt. “Required” is from an economic perspective, not a regulatory one.
Milliman - Economic Capital Modelling - pg. 4
Define available capital
The excess of the value of the company’s assets over the value of its liabilities on a realistic or market-consistent basis
Milliman - Economic Capital Modelling - pg. 4
What is a way to think of cost of capital?
Shareholders provide capital and expect to earn a reasonable return. This is the cost of capital.
Milliman - Economic Capital Modelling - pg. 7
What are the main concerns with holding too much or too little capital?
1) Shareholders expect to earn a return on the capital provided, so if capital is too high, the cost of insurance is too high (because cost of capital is incorporated in product pricing)
2) If capital is too low, risk of insolvency is high.
Milliman - Economic Capital Modelling - pg. 7
What are the weaknesses in formulaic approaches to the assessment of required capital?
1) No link between the capital requirement and the effectiveness of risk management
2) Cannot handle all types of risk
3) Don’t necessarily cope with changes in the financial environment
4) Typically don’t allow for benefits of diversification
Milliman - Economic Capital Modelling - pg. 8
What do SCR and MCR mean?
Both are capital requirements (solvency and minimum) in the Solvency II framework.
1) SCR: Capital required so that there is a 0.5% probability that assets will not meet liabilities during the following year.
2) MCR: Absolute minimum level of capital, calculated by a simple formula. Reaching the MCR would trigger urgent action from the regulator.
Milliman - Economic Capital Modelling - pg. 9
Why would regulators prefer internal models to calculate capital requirements instead of the regulator’s own standard approach?
Assuming the internal models are approved by the regulators, internal models…
1) encourage insurers to measure and manage their risks
2) are more flexible (changing with financial markets and the company’s evolution)
3) represent the insurer’s business more closely
Milliman - Economic Capital Modelling - pg. 9
What is underwriting risk?
Risks associated with the events covered by the insurance (death, earthquake, etc) and the specific processes associated with the conduct of the business (pricing risk, PHB risk, underwriting process risk, etc)
Milliman - Economic Capital Modelling - pg. 14
Operational risk can be defined in many ways. Sometimes, the sub-risks can be divided into what 2 groups?
1) Operational failure risk covers failure of people, processes, and technology
2) Operational strategic risk covers environmental factors like the appearance of a new competitor, a tax/regulation regime change, and natural disasters
Milliman - Economic Capital Modelling - pg. 15
What are the 3 key risk components for modelling?
1) Volatility risk: uncertainty in the frequency and severity of events
2) Uncertainty risk: risk of models deviating from reality due to errors, assumptions, simplifications, etc
3) Extreme events (calamity)
Milliman - Economic Capital Modelling - pg. 16
Which risk components are diversifiable and which are systematic?
1) Volatility risk: reduced by diversifying the portfolio
2) Uncertainty risk: systematic because it cannot be reduced by increasing portfolio size
3) Extreme events (calamity): systematic?
Milliman - Economic Capital Modelling - pg. 16
How can risks be measured?
By measuring the effect of the risk on a company’s earning or surplus. There are 4 approaches to do this.
1) scenario-based model
2) static-factor model
3) stochastic-factor model
4) covariance model
Milliman - Economic Capital Modelling - pg. 18
How does a scenario-based model work?
It can be deterministic or stochastic. Risk capital is calculated by measuring the impact of specific scenarios to the distribution of loss. (Correlation between risks must be taken into account when stochastic scenarios are generated.)
Milliman - Economic Capital Modelling - pg. 18
How is a scenario-based model different from a stress test?
Scenario-based models use scenarios that cover multiple risk drivers. Stress tests involve giving one shock to specific risk drivers.
Milliman - Economic Capital Modelling - pg. 18
How does a static factor model work?
It’s based on a linear combination of static risk factors multiplied by a company specific amount like the amount of an asset class or premium income.
Milliman - Economic Capital Modelling - pg. 18
How does a stochastic factor model work?
1) Identify relevant risk factors
2) Conduct a sensitivity analysis for each risk driver to measure delta, gamma, or a scenario vector
3) Model the joint distribution of risks
4) Aggregate the resulting loss across all risk types (leading to its stochastic distribution). Risk capital is determined by applying a risk measure (VaR or TVaR) to the company’s total losses.
Milliman - Economic Capital Modelling - pg. 18
How does a covariance model work?
It is a special case of a stochastic factor model with multi-normal distributions, first order sensitivities, and VaR as the risk measure.
Milliman - Economic Capital Modelling - pg. 19
How can volatility risk of underwriting risks be quantified?
The difference of liability amounts between best estimate and the TVaR under stochastic simulation
Milliman - Economic Capital Modelling - pg. 19
What are the types of mortality risk?
1) Mortality trend risk (affected by medical improvements and new diseases)
2) Catastrophic mortality risk
Milliman - Economic Capital Modelling - pg. 20
How do insurers mitigate catastrophic mortality risk?
They purchase reinsurance and/or issue mortality bonds. Investors invest the principal, and if no extreme event occurs over the investment period, they receive the principal with interest (paid with premiums from the mortality insurance). If an extreme event occurs, the investors lose part or all of their investment, and the insurer uses the funds to pay the high number of mortality claims.
Milliman - Economic Capital Modelling - pg. 20
How does a default model work?
Explicitly model the rates of default and recovery. Ex: assume number of defaults over a period of time follows a Poisson distribution with parameters estimated from historical data.
Milliman - Economic Capital Modelling - pg. 21
How does a credit mitigation model work?
It considers not only the risk of default, but also the risk that an investment will lose or gain value due to changes in the corporation’s credit rating.
Milliman - Economic Capital Modelling - pg. 21
How do changes in market prices affect an insurer’s liabilities?
Changing asset yields could affect…
1) the market value of liabilities through the discount rate
2) the amount and/or timing of future liability cashflows (like performance bonuses).
3) policyholder behaviour which also may change the amount and/or timing of future liability cashflows (ex: lapse)
Milliman - Economic Capital Modelling - pg. 22
What are replicating portfolios used for?
To measure market risk, an insurer needs to consider the effects on its assets and liabilities. Market values for assets are easy to obtain, but not for liabilities. Therefore, a replicating portfolio can be created to measure the market value of liabilities.
Milliman - Economic Capital Modelling - pg. 22
What are the 2 main ways to quantify operational risk?
1) Simple add-on model
2) Stochastic frequency-severity model
Milliman - Economic Capital Modelling - pg. 23
How does a simple add-on model work?
Combine the anticipated costs for the various operational risks. Assume a certain degree of correlation and a certain confidence level.
Milliman - Economic Capital Modelling - pg. 23
How does a stochastic frequency-severity model work?
Operational risk scenarios are identified from the bottom up. Each scenario includes a story and frequency and severity parameters.
Milliman - Economic Capital Modelling - pg. 23
Why might shareholders and regulators have a different opinion on whether an insurer uses VaR or TVaR as a risk measure?
Shareholders may consider VaR adequate because once the net worth has been exhausted, shareholders have lost their investment, so they may not care about the severity of further losses. For regulators, the magnitude of loss is significant because it determines losses to policyholders (which damage the reputation of the insurance industry and the regulator).
Milliman - Economic Capital Modelling - pg. 24
What is stochastic analysis?
An analysis done by projecting future cash flows based on multiple scenarios of which probability distribution is defined.
Milliman - Economic Capital Modelling - pg. 25
What is a stress test?
An analysis done by projecting future cash flow based on a set of particular scenarios that could occur in some extreme environments but for which occurrence probability is not specified.
Milliman - Economic Capital Modelling - pg. 26
What is stochastic analysis used for?
1) Stochastic analysis, where the occurrence probability is allocated to each scenario, is necessary to calculate economic capital, which is defined as an amount to cover future losses with a certain confidence level.
Milliman - Economic Capital Modelling - pg. 27
If it is difficult to determine a probability distribution for the occurrence of a risk, how can an insurer calculate the capital requirement?
Judgement call must be made. Assume an occurrence probability just for technical purpose of economic capital calculation. Ensure users understand its nature and the sensitivity to changes in the assumption.
Milliman - Economic Capital Modelling - pg. 27
Risk neutral
Milliman - Economic Capital Modelling - pg. 27
