LAM Full

E.2.8 - Committee of Sponsoring Organizations of the Treadway Commission (COSO) Definition of an Internal Control

"Broadly defined as a process, effected by an entity's board of directors, management and other personnel, designed to provide reasonable assurance regarding the achievement of objectives relating to operations, reporting, and compliance.

C.2.1 - First Fundamental Theorem of Asset Pricing

"Price must be given as the expected discounted future cash flows, where the expectation is taken with respect to risk-neutral measure"

A.1.6 - 5 Methods to Retire Debt Before Maturity (Corporate Bonds)

1. Call and refunding provisions 2. Sinking fund provisions 3. Maintenance and replacement funds (doesn't retire bonds) 4. Redemption through sale of assets (usually restricted) 5. Tender offers - issuer buys back bonds based on PV at CMT + fixed spreads 1-4 must be included in the indenture if applicable

A.2.3 - (Credit Derivatives) - 2 ways a protection seller can settle if a credit event impairs the reference entity bond

1. Cash Settlement: PS pays cash = Par - MV of Deliverable Obligations 2. Physical Settlement: PS pays notional in cash to PB, who delivers bonds to PS

A.1.12 - 3 Types of CMO Analysis

1. Cash flow analysis - test prepayment sensitivity, etc. 2. OAS analysis - compare classes of similar duration 3. Hedging - use key rate parameter or partial duration analysis

E.1 - AAE Principles of Actuarial Professionalism (5)

1. Integrity 2. Competence and Care 3. Compliance 4. Impartiality 5. Communication

C.1.1 - Two Key Classes of Non-Linear PRNGs

1. Inversive Congruential Generators (ICG) 2. Binary Shift Register Generators

D.1.2 - 6 Steps in GSAM's ALM & SAA Process

1. Investment Objectives and Constraints 2. Asset Universe and Assumptions 3. Liability CF & Replicating Portfolio 4. Risk Measures 5. Risk Return Trade-Offs 6. ALM & SAA

E.1 - IAA Principles of Actuarial Professionalism (3)

1. Knowledge and expertise 2. Values and behaviors 3. Professional Accountability

C.3.2 - State the 3 Formal Conditions of Linear Models

1. LM1 (Random Component) - Each component of Y is independent and normally distributed. The mean, u, of each component is allowed to differ, but they all have common variance. 2. LM2 (Systematic Component) - The covariates are combined to give the linear predictor, n 3. LM3 (Link Function) - The relationship between the random and systematic components is specified via a link function.

D.2.5 - List considerations when deciding between the general guidance approach and specific guidelines.

1. Level of cash flow 2. Asset risks inherent in the investment strategies 3. Investment advisor's expertise in the products your company sells 4. Management's comfort level with the investor advisor 5. Detail available from GL (including electronic availability)

A.2.1 - Steps to Find the Cheapest-to-Deliver (CTD) Bond

1. Identify all available bonds falling within maturity range 2. Calculate the price of each one at the underlying's YTM 3. The CTD bond is the bond that results in: Invoice Price - (Futures Price at Maturity * Conversion Factor) = 0 - Only one should satisfy above, rest should be negative.

E.2.6 - 6 Suggested Practices / Success Path for Overcoming Assumption Governance Challenges

1. Improve data availability and the assumption management process 2. Leverage existing resources or invest in new resources 3. Start early and focus on material risks 4. Increase comms between assumption owners and model owners 5. Reframe judgement and supplement documentation to meet third-party needs 6. Seek info and common industry benchmarks and practices

D.2.2 - 3 Major Sources of Hedging Error

1. Incorrect Duration 2. Projected Basis 3. Yield Beta

A.1.9 - 3 Factors that Increase a Floater's Price Volatility

1. Longer time between coupon reset dates (makes floater behave more like a fixed rate bond) 2. Decrease in market's required margin 3. Cap or floor is reached

A.2.3 - 6 Steps in the CDS Settlement Timeline

1. Suspected credit event occurs 2. Send request to DC to consider event 3. Compression cycle - reduces # of contracts that need to be settled 4. If hard credit event, proceed to auction; else for restructuring events only: -Publish maturity buckets -Triggering of CDS contracts -Deadline of movement option 5. Auction is held 6. Auction settlement

D.2.2 - 2 Extension of Basic CF Matching

1. Symmetric CF Matching -Borrow short-term money to meet liability -Invest in longer assets that will mature after the liability 2. Combination Matching (aka Horizon) -Duration-match portfolio and CF match initial years -Ensures short-term CF are met (e.g. 1st 5 years) -Reduces risk of non-parallel shifts -Con: increases cost of funding liability

A.1.12 - List issues in OAS Analysis

1. Term structure models that are too simple or inconsistent 2. Forward curve bias (but can exploit with derivatives) 3. Prepayment model accuracy (account for age, relative coupon, impairments, and burnout) 4. Securities created from collateral may be more valuable than collateral 5. Deal call risk (e.g. cleanup calls)

B.1.2 - 6 Methods to Improve Model Efficiency

1. Transfer Scenario Order - Identify only the worst scenarios within a larger set 2. Representative scenarios (4 types) - Select a subset of scenarios that represent the full set based on certain characteristics 3. Importance Sampling - Sample more scenarios in parts of the distribution that are most critical 4. Curve Fitting - Fit a underlying distribution 5. Cluster Modeling 6. Replicating Liabilities - Use optimization to determine a scaled subset of policies with similar characteristics to full inforce 2a. Modified Euclidean distance 2b. Relative present value distance 2c. Significance method 2d. Scenario cluster modeling

D.2.2 - Considerations When Applying Dedication Stragies (4)

1. Universe conditions (credit risk, embedded options, liquidity) 2. Optimization 3. Monitoring (periodic performance measurement) 4. Transaction costs (initial and rebalancing)

B.1.3 - Copula Risk Aggregation Approach

A Copula joins the marginals into the joint distribution Advantages: -More flexible - allows for skewness, non-linearity, and heavy-tailedness -Can allow for rich interactions between risks Disadvantages -Functional form and parameters have a fundamental impact on risk aggregation -Require estimation of the distributions for ALL underlying risk categories -Demanding Monte Carlo computation -Parameter estimation uncertainty -Challenging to communicate

B.1.2 - Importance Sampling Pros and Cons

Pros -Works best for options far out of the money Works well if: 1. Weights are designed to match CF pattern 2. Scenario ranking method is well-defined and closely correlated with CF drivers Cons -not better than significance method -Error increases as tail group size falls -Error increases as compression increases (true of all)

B.2.2 - Wilkie Model - Definition, Pros, and Cons

Multi-Variate Model -Project multiple economic variables -Vector autoregression for related AR(1) series Advantages -Designed for long-term application -Consistent projections of different variables Disadvantages -Designed for annual data only --> not good for hedging -Complexity of having multiple integrated models -Doesn't allow for major changes in the economy -Data mining problems -Inconsistent with the efficient market hypothesis

D.2.3 - 3 Key Rate Duration Propositions

P1: Effective duration = linear combination of KRDs = Sum(KRDs) P2: The KRDs of a portfolio = weighted sum of each bond's KRD Weight = price of bond / total portfolio value P3: A portfolio of zero-coupon bonds can be constructed to have the same value and interest rate exposure of the underling bond or portfolio

E.2.6 - Three Fundamental Concepts for Sustainable Assumption Governance

1. Guardrails 2. Risk-Based Framework and Monitoring System 3. Communication

B.2.2 - Advantages of Regime-Switching Lognormal Model

-Includes stochastic volatility -Fits equity-linked guarantee data well -Much fatter left tail than LN model -Puts more weight on extreme events

B.3.1 - Advantages of First Principles Modeling (LTC - active vs. disabled modeling)

-Internal assumption consistency -Refined assumption detail -Better benchmarking capability -A projection model that calculates paid claims and claim reserves on a more granular basis -A projection model that includes incidence of new claims and counts of existing claims

A.1.5 - Characteristics of Municipal Bonds

-Issued by states, local governments and other public entities -Usually purchased for their tax exempt status -Credit risk has increased in recent history -TCJA made municipals less appealing to corporate investors

C.2.2 - Issue with Large Alpha Values of Structured Indices in ESGs

-The source of the alpha needs to be understood, as it may be due to more than just diversification -The algorithm that shifts the portfolio make-up of the structured index may be "tuned" / overfitting to the ESG

A.2.2 - Describe Swap Terms -Trade date -Settlement date -Effective Date -Maturity Date -LIBOR "flat" -Swap spread

-Trade date: date parties commit to the swap -Settlement date: when parties actually exchange CFs -Effective Date: when swap starts accruing interest -Maturity Date: when swap stops accruing interest -LIBOR "flat": when there is no spread added to LIBOR -Swap spread = spread over N-year Treasury

B - Describe the Dynamic Hedging (Strategy)

-Use Black-Scholes equation to find replicating portfolio -Requires a stochastic interest rate assumption -Invest in replicating portfolio --> recalculate and adjust frequently

B - Describe the Actuarial Approach (Strategy)

-Use stochastic simulation to project liabilities -Use long-term fixed rate of interest to discount -Use risk measure to determine amount of capital needed to cover risk -Invest capital in risk free bonds

D.2.5 - Items to Include in an Insurer's Investment Policy (6) (IIISAA)

.1. Investment Objectives 2. Investment Constraints 3. Investment Committee 4. Scope 5. Authority from Board 6. Appendices

A.1.9 - 4 Portfolio Strategies Using Floaters

1. ALM - back short-term liabilities with floaters 2. Risk arbitrage strategies - buy floater with cheaper borrowed funds 3. Betting on changes in required margin 4. Swap arbitrage - enter a pay-fixed-for-floating swap that earns higher yield than comparable floater

D.2.5 - 4 Constraints on Asset Sales

1. Accounting - Stat accounting spreads gains and losses over life of original asset, realized gains are taxable 2. Embedded Value / Economic Value Added - Impact is usually minimal 3. ALM - Riding the yield curve 4. Credited Rates and Policyholder Equity - Selling a high-yielding bond can hurt crediting rates

B.1.1 Four Methods to Model Tail Behavior (LRM-130)

1. Adverse Quadrant Correlation - only estimate correlation in regions that are adverse 2. Adverse Period Correlation - only estimate correlation in time periods that are associated with adverse market conditions 3. Rolling Correlation - Gives a sense of the time dependency of correlation 4. Tail Dependency Analysis - Analyze the possible empirical joint distribution of a risk factor pair

D.2.2 - 3 ALM Applications of Swaps

1. Alter asset and liability CFs 2. Adjust the portfolio duration 3. Cheaper/easier alternative to using a package of forward contracts

C.1.1 - Four Common Variance-Reduction Methods (for Monte Carlo Simulations)

1. Antithetic-variable technique 2. Control variate technique 3. Stratified Sampling 4. Importance Sampling

E.2.4 - Four Phases of the Risk Management Control Cycle

1. Assess 2. Evaluate 3. Manage 4. Measure

D - Criteria for Specifying Asset Classes (5)

1. Asset classes should be relatively homogenous 2. Asset classes should be mutually exclusive 3. Classes should be diversifying (general rule: avoid pairwise corr. > 0.95) 4. All asset classes ~ world investable wealth 5. Each class should be able to maintain the portfolio's liquidity (should be able to buy / sell to rebalance)

B.2.1 - 4 Ways of Creating Provisions for EIAs

1. Buy options from third parties - form of reinsurance since all risk is passed to the seller of the option 2. Dynamic hedging - maintain a replicating portfolio 3. Actuarial Approach - perform stochastic analysis to develop quantile reserves 4. Ad hoc approach - use guesswork and/or actuarial judgement

D - 3 Types of Immunization

1. Classical Single-Period Immunization - Produces assured return for specific time horizon 2. Multiple Liability Immunization - Produces enough funds to pay all liabilities when due 3. Immunization for General Cash Flows - Given schedule of liabilities met by investment funds that are not fully available at the time the portfolio is constructed

D.1.2 - 3 Ways to Create an Integrated Risk Management Strategy

1. Clearly define a market risk budget 2. Evaluate economic objectives vs. insurance constraints 3. Determine how and where ALM fits into the overall risk management framework

E.2.3 - 6 Phases of an Actuarial Assignment Where Professional Judgement May be Required

1. Code of professional conduct duties 2. Choice of data 3. Dealing with missing or incomplete data 4. Choice of a model 5. Selection of model's key assumptions 6. Interpretation of model's outcome

E.2.4 - Five Components of Model Risk (CIIOR)

1. Conceptual Risk 2. Implementation Risk 3. Input Risk 4. Output Risk 5. Reporting Risk

E.2.1 - Three Groupings for Model Process Controls

1. Data Validation 2. Model Validation 3. Projection analytics/dynamics

E.2.1 - Four Basic Sources of Model Risk

1. Data limitations (in terms of availability and quality) 2. Estimation uncertainty or methodological flaws in model design 3. Calculation or coding error 4. Inappropriate use of a model

B.1.1 - Steps to Take Before Running a Cluster Model

1. Define location variables and their weights (Reserves, CSV, Prem, etc. - higher weight = higher priority) 2. Define a size variable (face, AV) (smaller policies get mapped) 3. Define segments that should NOT be mapped across 4. Specify a target number of clusters

A.1.3 - 4 Different Types of Securities Issued by US Treasury

1. Discount Securities (no coupons) - T-bills have maturity =< 1 year 2. Coupon securities -"Notes" have 2-10 yr. maturities (vast majority outstanding) -"Bonds" have maturities > 10 yrs. 3. Treasury Inflation-Protected Securities (TIPS) -Principal is inflation adjusted by CPI -Coupons = Fixed % of principal 4. Floating Rate Notes (FRNs) -Debuted in 2014 -2-year, fixed-principal notes that pay floating interest quarterly -Floating rate is based on 13-week t-bills

A.1.12 - 5 Evaluating Decisions to Buy Planned Amortization Classes (PAC)

1. Do I need CF stability? 2. Is the CF stability cheap? 3. Is it cheaper to hedge non-PACs with options? 4. Is the bond market range-bound? 5. What does current and implied volatility look like?

D.2.2 - Primary Bond Risk Factors (7)

1. Duration & Convexity 2. Key rate duration and PV distribution of CFs 3. Sector and quality percent 4. Sector duration and contribution 5. Quality (credit) spread duration contribution 6. Sector/coupon/maturity cell weights 7. Issuer exposure (manage event risk)

B.2.7 - Differences between EIAs and VAs

1. EIAs have shorter terms (~7 yrs. vs. 20-30 years) 2. EIA guarantee = call option on an index (VAs are puts) 3. EIAs are usually ITM at the maturity (VAs much less) 4. EIA sellers reinsure guarantee risk by buying call options (VA sellers assume more of guarantee risk) 5. EIAs are based on prices indexes (VAs use total return indexes)

C.2.1 - ESG Model Components (ESC-FO)

1. Equity returns 2. Sovereign interest rates 3. Corporate bond yields and returns 4. Foreign exchange rates 5. Other variables (inflation, GDP, unemployment)

E.2.6 - 5 Byproducts of Strategic Governance

1. Fostering communication among subject experts 2. Leveraging the best tech and data 3. Controlled, accurate implementation 4. Documentation and defensible processes 5. Satisfying management, regulator and auditor requirements

C.3.2 - Three Formal Conditions for GLMs

1. GLM1 (Random Component) - Each component of Y is independent and is from one of the exponential family of distributions 2. GLM2 (Systematic Component) - The covariates are combined to give the linear predictor n (n = X * B) 3. GLM3 (Link Function) - The relationship between the random and systematic components is specified via a link function, g, that is differentiable and monotonic

A.1.5 - Five Types of Municipal Bonds

1. General Obligation Bonds (secured by taxing powers) 2. Revenue bonds (backed by revenue from projects) 3. Hybrid and special bond securities (least risky) 4. Money market products (Notes, CP, VRDOs) 5. Municipal Derivatives (instruments form municipals)

D.2.2 - Strategies to Overcome the High Costs of Enhanced Indexing (5)

1. Lower cost enhancements - reduce trading costs and management fees 2. Issue selection enhancements - attempt to find undervalued securities 3. Yield curve positioning - find consistently mispriced maturities 4. Sector and quality positioning (2 forms) a. Tilt toward short corporates (high yield spread per unit of duration risk) b. Periodic over- or under-weighting of sectors and qualities 5. Call exposure positioning - e.g. underweight in callable bonds if you expect falling interest rates

D.2.2 - 3 Reasons to Use Indexing

1. Lower fees than managed accounts 2. Outperforming an index (after costs) is very difficult 3. Great diversification

C.1.5 - 3 Ways to Incorporate a Zero Lower Bond into Interest Rate Models

1. Make the volatility of interest rates proportional to the current interest rate 2. Impose a zero floor at each time step 3. Track the theoretical path of the interest rate separately from the lower bound

E.2.7 - DevOps Components (Microservices; Continuous Testing, Integration Delivery and Deployment; Infrastructure as Code; Telemetry; Continuous Feedback and Learning)p

1. Microservices - single-purpose libraries 2. Continuous Testing - unit tests are made for each unit of work within a microservce 3. Continuous Integration, Delivery & Deployment - should have small and singularly focused tasks to continuously deliver new features 4. Infrastructure as Code - concept that all aspects of the model and its configuration are in source control 5. Telemetry - Monitoring and logging the model by recording data on all mission-critical aspects of its behavior 6. Continuous Feedback and Learning - using an Agile approach of focusing on small and singularly focused tasks speeds up the feedback cycle

C.2.1 - Limitations of ESG Approaches

1. Model Risk 2. Sample error introduced by simulation 3. Long processing time for complex models 4. Challenge of reaching convergence 5. Requires a lot of data and expertise 6. Black box problem 7. May not adequately account for extreme events and regime changes

A.1.11 - 3 Model Duration Measures

1. Modified Duration - simplest but least reliable -V's reflect the interest rate shock but assumes CF are unchanged -Doesn't work well for passthroughs cash flows shift with interest rates 2. Cash flow duration - more accurate because it reflects cash flow shifts -Still not ideal since it reflects only one possible CF path per shock 3. Effective duration = uses a valuation model (most accurate) -V's are based on a monte carlo simulation of interest rates after each chock -Sensitive to assumptions in valuation model

D.2.5 - Items to Include in an Insurer's Liquidity Policy (5)

1. Objectives 2. Management Oversight 3. Liquidity measures and reports 4. Constraints 5. Written plan

A.1.14 - Key Items in a High-Yield Bond Offering (5)

1. Offering memorandum 2. Prospectus 3. Preliminary term sheet 4. Commitments by the underwriters 5. Industry overview and competitive position

A.1.14 - 3 Steps in Underwriting a High Yield Bond

1. Prepare the prospectus 2. Negotiate terms with investors 3. Syndication and allocation

D.2.5 - 6 Items to Include in an Insurer's ALM Policy (PROGGG)

1. Process 2. Reporting 3. Objectives 4. Ground Rules 5. Guidelines and Tolerances 6. Governance

D.2.2 - Strategies for Managing Against a Bond Mark Index, from the Least Tracking Error to Most

1. Pure Bond Index (full replication) 2. Enhanced indexing by primary risk factors 3. Enhanced indexing by small risk factor mismatches 4. Active management by larger risk factor mismatches 5. Full-blown active management

C.2.2 - 4 Determinants of Prepayment Rates (Agency Mortgage Passthroughs)

1. Refinancing - the most variable and interest-sensitive source of prepayments 2. Turnover (home sales) - not as interest-sensitive as a rate refi 3. Seasoning - refers to # of months since loan origination -Burnout: prepayments decline as past refi opportunities come and go 4. Defaults - 2 forms: -Strategic default: borrower "put option" to default when home value < loan -Any life event (job loss, etc.) that causes a borrow to default

D.1.3 - Five Best Practices in ALM

1. Secure senior management commitment 2. Ensure a clear assignment of roles and responsibilities 3. Leverage the CFT platform 4*. Select the most appropriate metric 5. Ensure a responsive and effective mitigation process *Metric = measurement for quantifying ALM risks (duration, convexity, VaR) **Must be relevant & actionable

C.2.1 - Key Steps in Parameterizing a RW ESG Model

1. Select the appropriate steady-state levels 2. Determine appropriate values for the initial conditions 3. Identify key parameterization targets for the application (stylized facts) 4. Control the expected mean reversion path

A.1.13 - 4 Tranches in CLOs

1. Senior Debt 2. Mezzanine Debt 3. Subordinated Debt 4. Equity

A.1.13 - Three Types of CMO Fees

1. Senior Management Fee -Paid after admin/trustee fees but before liability interest 2. Junior Management Fee -Paid after liability interest but before any payments to the equity class 3. Equity Class Incentive Fee -Like a hedge fund fee with a hurdle rate and fixed % of profits Senior + Junior fees ~ 30-35 bps -Can be lowered by arrangements called "side letters"

A.1.9 - 4 Spread Measures for Evaluating Floaters

1. Simple Margin (aka Spread for Life) - accounts for amortization of discount/prem in addition to quoted margin 2. Adjusted Simple Mean (aka Effective Margin) - #1 adjusted for "cost of carry" if borrowed funds are used 3. Adjusted Total Margin - #2 plus interest earned by investing difference in floater's par and the carry-adjusted price 4. Discount Margin - Calculates the average spread over the life of the floater Problem: none of these account for embedded options Solution: Monte Carlo

D.2.2 - Important Characteristics of Immunization (3)

1. Specified time horizon 2. Assured rate of return over a fixed holding period 3. Portfolio value at the horizon date is insulated from interest rate changes

B.2.4 - Black-Scholes Assumptions (5)

1. St follows Geometric Brownian motion (GBM) with constant variance 2. Frictionless market (no transaction costs, taxes) 3. Short selling allowed 4. Continuous trading 5. Interest rates are constant

D.1.6 - 3 Common ALM Methodologies

1. Static Methodologies (Dedication/CF Matching) 2. Dynamic Passive (Immunization/Indexing) 3. Dynamic Active (Active Mgmnt/Contingent Immunization)

D.2.1 - Quantitative return objectives with respective to SAA

AO: absolute asset returns ALM: asset returns net of liability growth

B.2.2 - Auto-Regressive Models vs. LN Model

Advantages Over LN: -Does not require I.I.D. variables -Mean reverting -Auto-correlation capture by AR -ARCH/GARCH produce volatility clustering Disadvantage: -Does NOT capture extreme values

A.1.7 - Qualities of Commercial Paper (& who uses it)

CP has: -low duration -high liquidity -generally high credit quality -Attractive place to park cash Money market funds held 23% of CP in 2019

B.2.2 - Lognormal Model Pros & Cons

Pros -Simple and tractable -Reasonable results over short time periods Cons -Not as good for long-term projections -Fails to capture extreme price movements -Does not allow for autocorrelation -Does not capture volatility clustering

C.1.1 - Stratified Sampling

Divides the distribution of a market variable into homogenous subgroups called stratum. Each stratum is sampled individually according to its probability. Steps: 1. Determine the interval (0-m) 2. Sample n number of v 3. Evaluate the function f at n random points uniformly distributed over stratum i and its average 4. Define the weights as proportional to the length of the interval 5. Compute the estimate f-hat 6. Repeat the steps above for J = 1, ..., N. The final estimate is the average value from step 5

C.2.1 - Define Economic Scenario Generator

ESG: A computer model of an economic environment that is used to produce simulations of the joint behavior of financial market values and economic variables Output: interest rates, equity returns, exchange rates, etc.

C.1.1. - Antithetic Variable Technique

Each random draw has two associated payoff calculations, the selection and the inverse. Example: If 99th percentile, u = .99 is generated, the opposite draw, u = .01 is additionally incorporated.

C.2.2 - Jensen's Inequality

For any random variable X, if f(X) is a convex function, then: E(f(x)) >= F(E(x))

C.1.1 - Control Variate Technique

Identify a simpler function, g(u), that has a closed-form solution. Can then be leveraged to estimate f(u). Most often used when an analytical solution of f(u) is unavailable.

D.2.5 - Describe how an investment actuary adds value

Investment Actuary = liaison between product actuaries and portfolio managers 1. Helps product actuaries/senior management (objectives, product design) 2. Develops policies for liquidity, ALM, investments, and derivatives 3. Provides product info to investment actuary 4. Helps set crediting rates 5. Coordinate cash levels with Treasury department 6. Advises portfolio manager on accounting/product constraints on specific trades 7. Allocates new securities to asset segments 8. Leads A/L modeling (scenario analysis, duration, hedging)

C.2.1 - Describe how ESGs different from other economic models

KEY DIFFERENCE: ESGs are NOT intended to be predictive: -ESGs produce scenarios that represent what could occur -No requirements to provide insight about why the economy works the way it does Example of other ESG models: -Econometric models -Forecasting models -Stochastic term structure models

A.1.1 - Relationship between linked IRR and time-weighted return

LIRR is a good proxy for TWR -Unless: large external cash flows (10%+ of AV) and/or volatile account growth

B.1.2 - Cluster Modeling Pros and Cons

Pros -Can be used in many modeling software's -Fairly low error at high levels of compression -Compression levels can be adjusted to improve accuracy -Minimal time required for ongoing use Cons -Compression may create noise in roll-forwards -Has potential for bias

B.1.2 - Curve Fitting Pros and Cons

Pros -Can use AMOOF (excel tool with 22 built in distributions) -Not sure? Cons -There is no way to target left tail -Requires significant amount of effort and expertise -May not be good distribution

C.1.1 - Monte Carlo Simulation, Pros and Cons

Pros -Convergence rate is independent of the dimensions of the integral -Multi-dimensional integration is theoretically feasible -Flexible and easy to accommodate complex models Cons -Slow convergence -Might not be useful to value American-style options

B.1.2 - Transfer Scenario Order Pros and Cons

Pros -Easy to understand, apply and audit -Low ongoing effort -Easy first step -Policy subset can be improved with advanced compression -More powerful for higher CTEs Cons -Only works for tail metrics -Error >= 0%

B.1.2 - Replicating Liabilities Pros and Cons

Pros -Produces low errors at high levels of compression -Errors did not vary by metric (CTE, VaR, etc.) -Ongoing maintenance is minimal Cons -Requires significant setup effort -Upfront learning curve -Produces extreme levels of compression (to get less, actuary would have to create manual workarounds) -May introduce bias

B.1.2 - Representative Scenarios Pros and Cons

Pros -Significance method produced least errors -Easy to understand and use (low ongoing effort) -No bias in the full distribution of results -Can be applied at wide range of compression levels -No additional scenarios need to be generated Cons -Not good for tail metrics -Requires considerable testing before implemeting

D.2.2 - Describe Contingent Immunization

Pursue active management as long as their is a positive safety margin. Safety Margin = Current Portfolio Value - Minimum Value Required for Immunization

A.1.13 - Example buyers for each tranche of a CLO

Senior: money center banks, endowments, pension funds Subordinated: mutual/hedge funds seeking diversification from BBB debt with similar/higher returns Equity: structures product experts (credit hedge funds)

C.1.1 - Importance Sampling

Similar to stratified sampling except we draw more samples from areas of high interest. Example: A very OTM option will have most simulations give a payoff of 0 - this results in a very uncertain estimator. By assigning higher weights to the regions of interest, importance sampling can increase accuracy.

C.2.2 - Key Takeaway from the Feynman-Kac Theorem

The Feynman-Kac Theorom states that price can be computed as a risk neutral discounted expected payoff.

A.1.1 - Liquidity and default risk in agency passthroughs.

There is NO default risk - guaranteed receipt of all principal and interest. -GNMA guarantee is based on full faith and credit of the US government -FHLMC/FNMA based on implicit government backing HIGHLY liquid (2nd only to Treasuries)

D.1.6 - Short/Long View

Two positions acting as a link between savers & producers: 1. Short - financial institutions provide financial assets to the household sector 2. Long - Financial institutions use the CF from the short position to purchase securities supplied by the corporate sector