1
Q
Define ERM
A
Process of systematically and comprehensively identifying critical risks, quantifying their impacts and implementing integrated strategies to maximize enterprise value
2
Q
7 Key aspects of ERM
A
- Process (continuous)
- Enterprise wide basis
- Focus on material risks
- Upside and downside: when it differs from expected
- Risks and correlatiosn must be quantified
- Strategies develop to avoid, mitigate or exploit risk factors
- Strategies are evaluated as a tradeoff between risk and return
3
Q
4 risks for an insurance company
A
- Financial
- Operational: related with the execution of the company’s business
- Strategic: : making the wrong or right strategic choice
- Insurance Hazard : risk taken intentionally with the goal of making profit
4
Q
4 Steps of the ERM process
A
- Diagnose (General environment, Industry, Firm): COnduct risk assessment to determine critical risks
- Analyse (Correlations and Critical risks): model risks that exceed a threshold
- Implement (Avoid, Reduction of occurence, Mitigate the impact, Eliminate, Retention): implement various activities to manage the risks
- Monitor: monitor actual outcomes of the plans implemented against expectations
5
Q
ERM risk management
A
ERM will help with important management functions and strategic decisions
- Planning growth (G)
- Valuing companies for M&A (V)
- Determining capital needs (C)
- Setting reinsurance strategies (R)
- Identifying sources of significant risk (I)
- Managing asset mix (A)
6
Q
ERM Risk models evaluation - 4 elements that differentiate the quality
A
- Reflects the relative improtance of various risks
- Includes dependencies
- Modelers have deep knowledge of the fundamentals of those risks
- Modelers have a trusted relationship with management
7
Q
ERM Risk models evaluation - Good vs Weak model
A
- Good model: Show as realistically as possible the risk and reward from a range of different strategies
- Weak model : May overstate and understate some risks
- Good model: Recognizes its own imperfections
- Extent and quality of the data
8
Q
4 Main Elements of ERM
A
- UW risk (day to day losses and cat losses)
- Reserve risk (Reserve runoff model to test for qualiy of fit)
- Asset risk (Model bonds, equities and foreing exchange rates)
- Dependencies/Correlations (macroeconomic condition (ESG will make other risks vary), across lines of business, extreme events - model with copulas)
9
Q
4 risks to consider under the UW risk
A
- Loss frequency and severity distributions (test quality of fit, understand remaining uncertainties, having a good modeler)
- Pricing risk
- Parameter risk (Estimation (estimate standard error to simulate different sets of parameters), Projection (trends changing in the future), Event and Systematic risks)
- Catastrophe model uncertainty
10
Q
ERM in setting capital requirements
A
- Based on default avoidance: 99% chance of not defaulting in next year (disadvantage:shareholders would be hurt at much lower loss amounts)
- Based to maximize franchise value (advantage: considers events at much lower loss amounts)
- Service renewals
- Not only survive a cat, but thrive in its aftermath
11
Q
Defne the four types of parameter risk
A
Estimation : misestimation of the parameters due to imperfect data
Projection: changes over time and the uncertainty in the projection of these changes
Event risk: major external event affecting your losses
Systematic: Risk non diversifying
12
Q
Three methods to make decisions under uncertainty?
A
- Deterministic : PV of CF estimated
- Risk Analysis: Distribution of outcomes is the output; most uncertainty is moved into the models of the critical inputs.
- Certainty equivalent: Uses the utility function on the outputs to formalize some of the risk assessment and apply consistently the firm’s corporate risk policy
13
Q
Why would it be necessary to use the certainty equivalent method?
A
Both managers and owners are interested in preserving franchise value and want an internal corporate policy to help make risk management decisions more objective, consistent, repeatable and transparent.
14
Q
Why would it not be necessary to use the certainty equivalent method?
A
Because it’s assumed that investors have a diversed portfolio and are not compensated for firm-specific risk but only for systematic risk the company should no be concerned with firm-specific risk.
15
Q
Three characteristics that affects the company’s risk tolerance
A
- Organization’s size
- Financial resources
- Ability and willingness to tolerate volatility
16
Q
Describe a utility curve
A
Used to describe the risk rolerance.
Reflects that the utility for an extra dollar of profit is less than the utility of the previous dollar and losing an additional dollar is more painful than he previous dollar that was lost.
Reflects the degree of our risk aversion
17
Q
Define certainty equivalent
A
Fixed amount that the firm is indifferent between taking the risky portfolio or the risk amount
18
Q
According to the modern portfolio theory how would you choose among possible portfolios?
A
The efficient frontier graph does not tell you at which point on the curve you shoud lie.
The utility function does it. You can calculate the CE by portfolio and look at which one has the higher certaitny equivalent.
19
Q
Define EVA
A
Economic Value Added = NPV - Cost of Capital
20
Q
3 Methods to monitor solvency by regulators
A
- Leverage ratios
- Risk based capital models
- Scenario testing
21
Q
Disadvantage of leverage ratios
A
- Same limits (thresholds) were applied for all types of business
- No distinguishing between LOB
22
Q
Advantages of RBC models
A
- Combines several risks (asset, credit, premium, reserve, catastrophe)
- Factor model so it varies with the quality and type of asset or LOB
- Can focus on long term viability (higher factors, AM best/S&P) or one year solvency (lower factors; regulatory models)
23
Q
Explain what scenario testing is
A
Test capital against a set of static (or stochastic more recenlty with associated projected financials) scenarios
24
Q
Define asset liability matching
A
Hedging interest rate risk by matching duration (matching duration of liabilities with assets or matching the cash flows by year)
25
Define asset liability management
It looks at current assets and liabilities as well as flows from future premiums of a going-coincern company.
26
4 optimal scenarios analyzed under ALM
1. Asset PF ( efficient frontier to find optimal portfolio ST vs LT and their related riskiness)
2. Known liabilities/CF (reinvestment risk or interest rate risk)
3. Liabilities and timing are variable (precise matching is not possible, must use a model that incorporates asset and liability fluctuations)
4. Going concern company, with CF from current operations (Can have +/-ve CF, must have an enterprise-wide model that includes premium, income, losses, cat and expenses)
27
Describe the asset-liability modeling approach ( 9 steps)
1. Models of asset, liabilities and business operations
2. Define risk metrics (income based or balance sheet based)
3. Define retun metrics (income based - earnings per quarter or balanche sheet based - ROE)
4. Time horizon (single year vs multiperiods - harder to implement and must include correlations)
5. Consider relevant constraints (limits by regulators, investment policy of the company,RBC score threshold)
6. Run simulations that incorporate different UW, Reinsurance and Investment strategies
7. Efficient frontier graph
8. Reflect that liabilities (future loss reserves) can be modified by reinsurance or UW strategies
9. Review situations where prefered portfolio performed poorly
28
3 ways to think of the value of reinsurance
1. Provides stability
2. Frees up capital
4. Adds to the value of the firm
29
Net cost of reinsurance
Ceded premiums (net of ceding commission) less the recovered losses (net of reinstatement premium)
30
Metrics to compare value of reinsurance
1. Reinsurance loss ratio (expected recoveries / ceded premiums)
2. Net UW profit (premiums (net of reinsurance) less expected losses (net of expected recoveries))
3. Differences between distributions at different percentiles
4. Space Needle view
5. Cost benefit diagram
6. Financial ratios
7. Efficient frontier (eg: earnings against probability of making the plan)
31
Explain Cost benefit diagram
At a given percentile graph the expected loss at this percentile against the cost of reinsurance for many scenarios to see which one performs better. Do it at many percentiles too see the overall performance
32
Why does reinsurance reduced the required capital
Because a business with stable results will require less capital. Therefore if a reinsurance program reduces the volatility of the results the amount of capital required is reduced.
33
Two methods to analyze if the reinsurance program has a dollar benefit and how to select between more than one
1. Change in cost of capital must be greater than the net cost of reinsurance
2. ROE (net cost of reinsurance/change in capital required) must be less than the firm's cost of capital b/c here we are investing "negative money".
34
Reinsurance as capital : 2 ways to determine the change in capital from the reinsurance structure
1. Theoretical models: Compare options with their ROE (want smaller ROE below the company's cost of capital).
ROE = change in earnings/change in required capital
Change in earnings = Expected earnings of this option - Expected earnings of current option.
Expected earnings = Net P - Net L - Expenses (net of reinsurance)
2.Pactical models: Use capital requirements from rating agencies
+Easy to implement b/c no need to simulate losses
-Not as accurate as theoretical model
35
How to calculate TVaR of earnigns
Net Premiums - Expenses - TVaR of losses
36
Advantages of Economic capital
1. Unifying measure for all the risks across an organization
2. More meaningful to management than formulaic RBC or capital adequacy ratios
3. Forces firm to quantify the risks it faces into a probability distribution
4. Provides a framework to set capital for the firm as a whole and for individual business units
37
3 types of risk measures
1. Moment based
2. Tail based
3. Probability transform
38
Examples of moment based risk measures
- mean
- variance
- standard deviation (preferred to variance b/c same unit base)
- skewness
- Ye^(cY/EY) captures all moments but need a maximum possible loss
39
Examples of tail based measures
- Value at risk (VaR = Loss at a given percentile)
- Tail value at risk (TVaR = Average of loss above given percentile)
- Excess tail value at risk (XTVaR = TVaR - mean)
- Expected policyholder deficit (EDP=(1-alpha)*(TVaR-VaR))
- Default put value : Value of insuring against default, can use EDP with a probability transform distribution
40
Disadvantage of tail based moments
Does not consider small losses that would still affect the company badly.
Events not in the tail are still important
41
Which risk measures to use?
1. If use TVaR, use it at a lower percentile to include all risk below the high percentiles
2. And try to get close to market value (WTVaR with the minimum entropy function and the exponential moment)
42
Define marginal allocation
An allocation is marginal if the change to the company's risk measure from a small change in a single business unit is attributed to that business unit
-Only one co measure can be marginal per risk measure
43
Define suitable allocation
A growing business unit with an higher than average profit/risk will increase the profit/risk for the company
44
Why is allocating capital arbitrary and artificial
Arbitrary because different risk measures gives different allocations (so different answers)
Artificial because the business unit has access to the entire capital of the firm, not just a portion
45
4 steps mechanism for decision making with an internal risk model
1. Starts with an aggregate loss distribution including many risks
2. Quantifies the impact of the possible aggregate loss outcomes on the corporation
3. Assigns a cost to each amount of impact
4. Attributes the costs back to the risks sources
Need the firm to know how much risk it's willing to tolerate, how much reward it's willing to give up for a reduction in risk
46
The amount of capital an insurance company holds is a function of a number of things such as:
1. Customer reaction (security vs price)
2. Capital requirements of rating agencies
3. Comparative profitability of new and renewal business
47
How to include accumulated risk when measuring the impact of reinsurance
As-if loss reserves: Reserves that would exist at the beginning of the accident year if that business had been written in a steady state in all prior years.
Combine current accident year reserve with the as-if loss reserves to obtain a proxy for capital consumed by an accident year over time.
48
2 advantages of as-if loss reserves
- Can measure the impact of accumulated risk caused by correlated risk factors across accident years
- Reinsurance being considered can be applied to the accident year and as-if loss reserves, providing a more valid measure of the impact of reinsurance on accumulated risk and on capital absorbed over the full life of the AY.
49
Define corporate risk tolerance
Refers to the organization's size, financial resources, ability and willingness to tolerate volatility
50
Describe how return on risk-adjusted capital (RAROC) is determined and used to determine if an activity is worth pursuing
First allocate risk capital to portfolio elements.
Then multiply the allocated risk capital by a hurdle rate to determine RAROC for each element.
Calculate the EVA bu substracting the RAROC from the NPV of the activity's CF.
If EVA is positive it should be pursued
51
Describe how to use cost benefit analysis if the risk capital is allocated or if the cost of capital is allocated.
If risk capital is allocated: Compare benefit (reduction in required capital) with cost of implementing.
If cost of capital is allocated: Look if positive incremental EVA.
52
Two disadvantages of using standard deviation to measure risk
1. Favorable events are treated the same way as unfavorable ones.
2. As quadratic measure it may not adequatly capture market attitudes
53
Disadvantages of TVaR
1. Does not capture events below that percentile
| 2. Linear in the tail
54
Describe WTVaR
Once you calculated transformed probabilities to put more weight on bad events you can calculate the TVaR on that distribution which won't be linear in the tail
55
Describe how the ocmpany can allocate the cost of capital
Set the minimum profit target of a business unit equal to the value of its right to call upon the capital of the firm.
The excess of the unit's profits over this cost of capital is added value for the firm.
We are allocating the overall firm value to each business unit
56
Two disadvantages of leverage ratios
1. Dont distinguish between business classes
| 2. Dont include other risks than underwriting
57
Compare RBC models to leverage ratios
1. They combine different aspects of risk into one number
2. They include 4 types of risk (vs only 1 UW):
- credit
- invested assets
- premium
- reserve
3. Multiply a factor by accounting values and the type and quality of assets are considered when determining the factor to be applied
58
In what type of accounting do the assets provide hedging against liabilities
Economic accounting
59
Explain how bonds are measured in the three different accounting systems
Statutory : Amortized
GAAP: Marked to market
Economic: Marked to market
60
Explain how liabilities are measured in the three different accounting systems
Statutory: Undiscounted
GAAP: Undiscounted
Economic: Discounted
61
4 Considerations on implementing internal risk models
1. Startup: Staff and scope
2. IRM parameter development
3. Implementation
4. Integration and maintenance
62
6 Elements in Startup: Staff and scope
Organization chart: leader with fairness and balance
Functions represented: departments included
Resource commitment
Critical roles and responsibilities: Controls of inputs/outputs and analyses
Purpose
Scope
63
Important inputs to the model
```
Expected loss ratio
Distribution of the loss ratio
Expected premium
Correlations
Economic scenario generator
```
64
4 Considerations in IRM Parameter development
- Modeling software: capabilities of the IRM team
- Parameter development: Include all departments, develop systematic way to include expert opinion
- Correlations: Owned at high level b/c has significant impact on capital and crosses lines of business
- Validate: Test over extended period and train interested parties so that they understand the statistics
65
4 Considerations in Implementation
1. Priority setting: Top management so that they are aware and raise their concerns rapidly
2. Communications: Regular and broad audience
3. Pilot testing: Effective preparation for the magnitude of the change
4. Education: Train leadership to have similar base level of understanding
66
3 Considerations in Integration and Maintenance
1. Cycle: integrate into regular planning
2. Updating : Max 2 times per year for major changes
3. Controls: Maintain centralized control of inputs
67
Why is it important to model parameter risk
Because for large companies the process risk gets diversified and we are left with parameter risk only.
68
3 risks that need to be considered relating with parameter uncertainty
Estimation risk
Projection risk
Model risk
69
Explain projection risk
We forecast future trend using historical averages when historical costs are not fully known.
Must include two things:
1. Superimposed inflation that varies with general inflation in the model
2. Trend as a time series instead of simple trend model
70
Explain estimation risk
Risk of misestimation of the parameters.
The parameters are based on MLE which give the highest probability of generating the actual data set.
Must include the possibility that we are uncertain about the selected parameters.
We do this by assuming the parameters follow a joint lognormal distribution so that they don't generate negative parameters and are heavy tailed (if joint normal distribution)
71
Explain model risk
Risk that we did not select the correct distribution fo our parameters because different sets/distributions of parameters yield the same likelihood.
Must simulate which distribution will be used
After that simulate the parameters that will be used
72
2 recommendations when using projection models
1. Ensure the model is structuraly consistent with the underlying process
2. Make sure it gives a realistic spread of potential outcomes (vs best estimate; not used the same way)
73
Why is the pearson's correlation not useful for the loss distribution for the company as a whole
Non symmetric distribution
| Have heavy tails
74
Formula for Kendall's T and 2 advantages
+ Relies on the order (instead of the values)
+ It will be the same for a copula no matter what the underlying distributions are
T = C-D/(C+D); compares the pairs to see if both variables are higher than two others to see if dependent
75
Describe a copula
Analyzes the dependency between two variables by just looking at the percentiles of the outcomes.
76
Describe the dependencies for the four copulas
1. Frank's
- symmetric
- light tails
2. Gumbel
- asymmetric
- more weight in right tail (more than frank)
3. Heavy right tail
- asymmetric
- high correlation in right tail
4. Normal
- symmetric
- higher density than frank
77
Describe the two tail concentration functions and their formulas
Left tail concentration function: The probability that U is small given that V is small.
L(z) = C(z,z)/z
Right tail concentration function: The probability that U is large given that V is large.
R(z) = (1 -2z + C(z,z))/(1-z)
78
What does J(z) represents
How does T build up from zero (when no data points are available), up to T when all data points are considered.
79
Define operational risk
The risk of loss resulting from inadequate or failed internal processes, people and systems or from external events
80
Give 7 categories of operational risks
Internal fraud (theft, claims flasification)
External fraud (false application, claims fraud)
Clients, products and business practices (client privacy, bad faith)
Employment and workplace safety (repetitive stress, discrimination)
Business disruption and system failures (processing center downtime, system interuptions, flood)
Damage to physical assets (damages to office, own auto fleets)
Execution, delivery and process management (policy processing, claim payment errors)
81
What is cycle management?
Prudent management of underwriting capacity as market pricing fluctuates in the underwriting cycle.
82
3 questions to ask relating to cycle management
- Does the company have a proactive cycle management strategy?
- Does the company know where in the cycle the market stands at any given time?
- Are underwriters make decisions that are consistent with the strategy and the market position?
83
4 characteristics to judge from a system performance perspective
Stable
Affordable
Reliable
Available
84
Performance improvement can be achieve if we focus on these 4 areas
- Intellectual property: must be retained, stay present in core markets and maintain a consistent pattern of investments
- Underwriters incentives: making the plan should be related to how well it supported the portfolio goals and underwriters must be aware that they may need to stop writing business
- Market overreaction : know that industry tend to overreact with prices; the more capital will help
- Owner education: they must understand that under cycle management some financial figures will be out of step with market; premiums will drop and expenses will jump
85
Explain what is the agency theory
Owners and managers may not have incentives that are aligned.
Managers may be too risk averse if they are compensated with stocks.
Managers may be too agressive if they are compensated with the firm's value in 5 years.
86
5 general operational risks
```
Pension funding
IT failure
Other HR
Reputational
Lawsuits
```
87
Define control self assessment
A process through which interna control effectiveness is examined and assessed. The objective is to provide reasonable assurance that all business objectives will be met.
88
Give examples of objectives of internal controls
- Reliability and integrity of information
- Compliance with policies, laws and regulations
- Safeguarding of assets
- Economical and efficient use of resources
89
Define key risk indicators, its goal and advantge
Broad categories of measures that monitor the activities and status of the control environment of an operational risk category.
Goal is to keep the risk management process dynamic and risk profiles current.
+Measured daily
+Forward looking
90
4 possible insurer's KRIs
1. Production: retention, avg premium, closing
2. Internal controls: audit results, audit frequency
3. Staffing: employee turnover, premium per employee, trianing budget
4. Claims: frequency, severity
91
Define six sigma
Tolerances for output quality are plus or minus three standard deviations from the mean.
Applied for existing process improvement and predictive design:
- UW: expo data verification, expo data capture and classification
- Claims: coverage verification, ALEA, use of outside counsel, initial case reserve
- Reinsurance: treaty claim reporting, coverage verification, reinsurance recoverables, letters of credit
+Useful when there is high volume processing
92
How to model operational risk
Identify exposure base (payroll, head count, policy count)
Measure exposure level by B unit and operational risk
Estimate loss potential per unit of expo by risk
Combine loss freq and sev by business units
Estimate the impact of mitigation, process improvements or risk transfer on the B unit loss distributions.
- Little data
93
Define UW cycle
Recurring increase and decrease of prices and profits.
94
Describe the evolution of insurance business (4 phases)
1. Emergence
- data is thin
- demand grows quickly
- leads to solvency crisis where weaker competitors are forced out
- profitability follows leading to new entrants
- this is classic UW cycle
2. Control
- stabilization
- government help to set maximum premiums
- rating bureaus or insurance departments help maintain stability
3. Breakdown
- changes
- new competitors take business away
4. Reorganization
- returns to phase 1
95
4 theories on the UW cycle
1. Institutional factors: lag between notice of price and inadequacy and inability to react rapidly
2. Competition: some will underestimate losses and charge below so results in downward pressure on whole group
3. Supply and Demand, Capacity constraints and shocks: losses that reduce capital will make some clients leave and capital cannot be replaced quickly
4. Economic linkages: investment income, cost of capital, expected losses (inflation, unemployment, etc) and price of risk
96
How to model the uw cycle
Soft Approaches:
1. Scenarios: written statement describing possible future states to determine company's reaction
2. Delphi method: gather expert opinion without biasing the group to the opinion of the most senior persons
3. Competitor analysis: follow some metrics and normal vs abnormal statistics will become evident
Technical modeling:
AR(2) or AR(3) on combined ratios for example
```
Behavioral modeling (Econometric):
Structural insight and statistical validity mix the two types of modeling
```
97
What increases the supply at a given price
- new entrants
| - technological changes
98
What decreases the supply at given price
- higher capital requirements
| - shock forcing out weaker insurers
99
Define strategic risk
Intentional risk taking to achieve an organization's goals.
The risk to earnings or capital arising from adverse business decisions or improper implementation of those decisions.
100
Define strategy
A strategy is a long term series of actions designed to take a company from its current state to its desired future state, while maintaining a competitive advantage
101
List 7 categories of strategic risk
1. Industry
2. Technology
3. Brand
4. Competitor
5. Customers
6. Project
7. Stagnation
102
What is a key part to effective strategic management?
Scenario planning
103
Key characteristics of scenario planning for strategic management
1. Use a limited number, each of which have different elements interacting with each other under different conditions
2. Test for internal consistency and plausability
3. Explore joint impact of several variables (effect of more than one change at a time)
4. Change several variables at one time to capture new states after major shocks
5. Scenarios will include subjective interpretation of factors, and thus cannot be explicitly modeled
104
10 key steps in scenario planning process
1. Define scope (time frame, geography, market segment)
2. Identify major stakeholders (employees, owners,customers, competitors, suppliers)
3. Identify basic trends
4. Identify key uncertainties
5. Construct initial scenario themes
6. Check for consistency and plausibility
7. Develop learning scenarios
8. Identify research needs
9. Develop quantitative methods
10. Evolve toward decision scenarios (review scenarios to converge to test scenarios and generate new ideas)
105
2 advantages of scenario planning
1. Company thinks through responses beforehand and agree on best responses ahead of time. Saves time during crisis
2. Operational inertia is reduced, flexibility is built into the system
106
How does advanced scenario planning works?
It includes responses based on a set of rules.
Steps:
- Define goals we want to achieve
- Determine essential environmental variables
- Define actions rules
After simulation we calculate the average return and volatility by set of rules and select a strategy that meets our goals the best.
107
Waht is Agent-based modeling?
Model how the company will respond to different market states, because the market is not static.
- Program agents (other parties in the system)
- Determine a set of rules for each agent to show how they react
- Run simulation
+Have emergent properties that describe the behavior of the agents
Exam 7
flashcards
Decks in class (4)
# Cards
