... connection with the return on capital, one can determin

1
Principles-based Reserves and RBCIowa
Actuaries Club February 28 2006David E.
NeveCo-chair Academy Life Reserves Work Group

2
Objectives of this Session

• 1. Define Principles-based
• 2. Discuss Academys Principles-based
  initiative (PBA)
• 3. Provide Summary and Update of PBA to Life
  Products
• 4. Briefly Summarize ACLIs Interim Solution
• 5. Briefly Describe work underway to develop
  new mortality tables.

3
Definition of Principles-based

• 1. Captures all of the material financial
  risks benefits and guarantees associated with
  the contracts including the tail risk and the
  funding of the risks.
• 2. Utilizes risk analysis and risk management
  techniques to quantify the risks. This may
  include stochastic models or other means of
  analysis that properly reflect the risks of the
  underlying contracts.

4
Definition of Principles-based

• 3. Incorporates assumptions and methods that
  are consistent with but not necessarily
  identical to those utilized within the companys
  overall risk assessment process.
• 4. Permits the use of company experience to
  establish assumptions for risks over which the
  company has some degree of control or influence.
• 5. Provides for the use of assumptions set on
  a prudent best estimate basis that contain an
  appropriate level of conservatism when viewed in
  the aggregate.

5
Definition of Principles-based

• In contrast a rules-based approach
• Relies on a static formula that may not capture
  all of the risks of the contract.
• Uses prescribed valuation assumptions that are
  the same across all companies regardless of
  differences in the risk profile of companies.

6
Observations of moving to a Principles-based
approach (PBA)

• 1. Is consistent with the global trend toward
  Enterprise Risk Management
• 2. Relies more on actuarial judgment
• 3. Requires more sophisticated tools
• 4. Requires that a stronger regulatory
  governance process be in place including
  independent review

7
Why Do We Need PBR

• 1. Current valuation system is broken we have
  been using band-aids as new products come out
• 2. Reserves are too high for some products
  too low on others
• 3. Capital requirements and reserves need to
  take into account actual risks of products and
  company

8
Evolution

• Asset Adequacy Testing
• Equity Indexed Annuity Regulation
• C-3 Phase I
• C-3 Phase II

9

• Academys Principles-based Initiative

Governance Issues and Overall direction SVL II
Work Group Reserve work groups Variable
Annuities VACARVM Life Products
LRWG Other Annuities ARWG Coming soon
LTC RBC work groups Fixed Annuities C3 Phase
I Variable Annuities C3 Phase II Life
Products C3 Phase III
10
Educational Opportunities

• 1. Webcasts Academy has committed to a one
  hour webcast after each quarterly NAIC meeting.
  Format presentations ask questions.
• 2. Academy website www.actuary.org/risk.asp
• Gives overview of projects
• Provides links to documents Seminars
• 3. Seminars

11
Life Reserve Work GroupOverview and Update
12

• LRWG Charge and Scope

• Charge
• Develop a proposal for a new Principles-based
  statutory reserve method for life products
• Coordinate with C3 Phase III work group (which is
  working on RBC requirements for life products)
• Scope
• Initially scope was limited to UL
• Now scope is all life products

13
Basic Framework

• Based on Gross Premium Reserve (GPR)
• Reserve PV of future benefits and expense
  (excluding FIT) less PV of future gross premiums

• Reserve assumptions will be determined for all
  material risks (mortality interest expenses
  lapse premium levels etc.)

• Reserve assumptions will include a margin for
  adverse deviation (not best estimates)

• Discount rates will be pre-tax

14
Basic Framework (cont)

• Reserve is the greater of
• A deterministic seriatim single scenario
  reserve calculation

2. A stochastically derived reserve (if needed)
using a prescribed CTE level
Since the stochastic reserve is done in the
aggregate risk offsets between contracts are
recognized (but limited).
15
Basic Framework (cont)

• Deterministic Reserve
• Uses a single set of assumptions that is aligned
  with economic reality yet still provides an
  appropriate level of conservatism
• Is not designed to capture tail risk
• Is subject to a cash surrender value floor on a
  contract by contract basis

16
Basic Framework (cont)

• Stochastic Reserve
• Multiple scenarios will be defined to properly
  capture the tail risk of the contract (risks
  that have high impact but low probability)
• Will use a CTE (conditional tail expectation)
  level that is set by regulators such as 65 CTE
• Current thinking is that only interest rate
  movements will be modeled stochastically

17
Basic Framework (cont)

• Prudent Best Estimate Assumptions
• Assumptions will be based on prudent best
  estimates that include a provision for adverse
  deviation
• Definition Conservative end of actuaries best
  estimate confidence interval
• Since actuarial judgment is involved will need
  to set limits and controls on setting assumptions

18
Basic Framework (cont)

• Asset Model Needed to Project Cash Flows
• Needed for both Deterministic and Stochastic
  Reserve
• Asset Model is used to determine
• Discount rates for GPR
• Earned rates for surrender benefits
• Discount rates for GPR calculation
• Based on projected portfolio rates in each year
• New money treasury rates will be prescribed for
  Deterministic Reserve modeled for Stochastic

19
Basic Framework (cont)

• Principles-based versus Asset Adequacy Analysis
• Both involve more actuarial judgment than
  current rules-based valuation approach
• Asset adequacy analysis has few limits and
  controls actuary has a high degree of
  discretion in setting assumptions
• In contrast the Principles-based approach
  will have controls caps and limits placed
  throughout the framework

20

• Challenges

• Establishing assumption margins (on each
  assumption and in the aggregate)
• Difficulty of projecting future premium levels
  for UL
• Criteria to require Stochastic Reserve
• Treatment of non-guaranteed elements
• Impact on taxes (tax deductibility 7702 issues)
  

21

• Major Issues Under Discussion

• Discount Rate
• Including Federal Income Taxes
• Allowing full aggregation
• Gross Premium Valuation vs. PV of greatest
  accumulated deficiency (VACARVM approach)
• Stochastic Reserve exclusion
• Small Company concerns

22

• Timeline and Deliverables

• LHATF exposed the proposal for comment at the
  December 1 2005 LHATF Winter meeting.
• During 2006 LRWG will assist LHATF in finalizing
  the details of the proposal (and will assist the
  C3 Phase III work group to finalize the RBC
  proposal).
• Goal is for LHATF to approve final draft of
  reserve proposal at December 2006 LHATF meeting
  and sends to A Committee for approval
• Begin state-by-state adoption in 2007.

23

• Key Issues Needing Regulatory Attention

• Development and implementation of an acceptable
  governance process
• Make changes to SVL that enable the
  Principles-based approach to be implemented by
  Model Regulation and/or Actuarial Guidelines
• Determine specific limits and controls on reserve
  assumptions and margins
• Decide if new approach will be applied to inforce
  contracts or only applied prospectively.

24
Determining Assumption Margins
25
Considerations for setting assumption margins
under a principles-based system

• 1. This is a major issue!
• 2. Regulators need to determine the balance
  between prescribed standards and actuarial
  judgment.
• 3. The LRWG believes there are several
  reasons why different margins are justified
  compared to current formulaic approach
• 4. The LRWG has developed a tool that
  provides a quantitative comparison of the
  aggregate impact of all assumption margins on the
  reserve

26
Reasons for different assumption margins under a
Principles-based approach

• 1. Assumptions reflect risk characteristics of
  each company no need to establish an
  industry-based margin to cover uncertainties
  between companies.
• 2. Assumptions are not locked-in at issue
  less need for a provision for adverse deviation
  since assumptions can be revised in the future
• 3. Implicit margins are already built into
  the methodology
• Blending to an industry mortality table if
  experience not fully credible
• Removing mortality improvement is a margin
• Cash value floor is a margin

27
Possible approach to compare aggregate impact of
all assumption margins
The LRWG is exploring is the use of a number we
are calling Z to provide for the quantitative
comparison of the aggregate impact of all
assumption margins. It is defined as follows
Z Reserve held - Best estimate liability
Present value of capital
requirement Z represents the amount by which
the pre-tax return on capital is expected to
exceed the return on invested assets
ROC Z i (pre-tax)
28
Possible approach to compare aggregate impact of
all assumption margins

• Given this connection with the return on capital
  one can determine whether the aggregate impact of
  all margins are within a reasonable range.
• For these illustrations the level of capital was
  set equal to 100 of claims plus 5 of the
  reserve.
• Z could be used as a disclosure item to compare
  the aggregate impact of all assumption margins.
  

29
Modeling Results20-year Level Premium Term
Product
30
Modeling Results20-year level premium term
product

• Initial results were presented to LHATF in
  December
• Overall reserves were lower than current
  formulaic reserves
• However concerns were expressed about the
  appropriateness of the high reserve levels in the
  early durations (higher than current formulaic)
• Upon review the impact was due to high
  assumption margins that were used in early
  years impact of high margins has large effect on
  PV of benefits but not PV of premiums.
• Conclusion overall margin levels must be
  carefully considered under the PBA especially
  for mortality (e.g. the loadings used for 2001
  CSO may be too high.).

31
Recap of Results from December LHATFExcerpts
from9 of December term presentation
32
Modeling Results20-year level premium term
product

• Decided that multiple scenarios using different
  assumption margins needed to be modeled
• Changes Made in Model since December
• Policy terminates end of 20th year (removed
  option to renew)
• Increased the premium level
• Assumed mortality fully credible (no blending)
• Modeled different assumption margins (primarily
  mortality margins)
• Formulaic Reserves updated to use 2001 CSO

33
20 Year Term Product Description
Plan of Insurance 20 Year Level
Term Guaranteed Premiums No Renewal Option
after 20 yrs. Gender/Issue Ages Male 45 and
65 Risk Class Best Non Smoker Class
(1) Reflecting capital of 100 of claims and 5
of reserves. Reserves using PBE assumptions
34
Market Perspective Premium ComparisonIssue Age
45 Best Class Annual Premiums for 1000000
Face Amount
35
Market Perspective Premium ComparisonIssue Age
65 Best Class Annual Premiums for 1000000
Face Amount
36
Five Margin Levels

Level 1 Deterministic interest scenario 2001
CSO mortality margins 30 lower lapse
rates Level 2 Same as level 1 but mortality
margin of 9.375 deaths per 1000 divided by
ex Level 3 Same as level 1 but mortality
margin of 3.5 deaths per 1000 divided by ex
Level 4 Deterministic interest scenario
mortality margin of 3.2 no other margins Level
5 Deterministic interest scenario mortality
margins of 2.1 10 lower lapse rates Best
Estimate reserve (no margins) is also shown
37

• Observations
• Level 4 and Level 5 margins give a near zero
  reserve at time 0 which is close to a no gain
  or loss at issue scenario
• The reserve at the end of the first year always
  decreases from time 0 due to acquisition
  expenses
• But the reserve is not forced to be zero at the
  end of the first year since there is no FPT
  adjustment as under the current formulaic
• Cash value floor would come into play (reserve is
  negative) in early durations for Level 3 4 and
  5.
• Deterministic reserve is about the same as the
  stochastic reserve.

38

• Observations (cont.)
• Current formulaic reserves start with small Z
  but then Z gets very large due to impact of
  mortality margin on PV of benefits and net
  premiums.
• Level 4 and Level 5 margins produce a Z value
  close to 4 consistent with a 10 IRR assumption
  (that is 4 over investment return)
• Levels 1 2 and 3 margins have significantly
  higher Z values (in excess of 20).

39
20 Year Term ExamplesDeterministic Terminal
Reserves at Different Margin Levels Male 45
Best Class 1000000 Annual Premium of
1415.00.
40
20 Year Term ExamplesDeterministic Terminal
Reserves at Different Margin Levels Male 45
Best Class 1000000 Annual Premium of
1415.00.
41
20 Year Term ExamplesComparison of Z Levels and
Deterministic Reserve Margins Male 45 Best
Class 1000000 Annual Premium of 1415.00.
42
20 Year Term ExamplesDeterministic Terminal
Reserves at Different Margin Levels Male 65
Best Class 1000000 Annual Premium of
11875.00.
43
20 Year Term ExamplesDeterministic Terminal
Reserves at Different Margin Levels Male 65
Best Class 1000000 Annual Premium of
1415.00.
44
20 Year Term ExamplesComparison of Z Levels and
Deterministic Reserve Margins Male 65 Best
Class 1000000 Annual Premium of 11875.00.
45
20 Year Term ExamplesModel Office Reserve
Levels Aged 20 Years

• Current Formulaic Reserve 121301
• Comparative Deterministic Reserves
• Level 2 50564 (42 of Formulaic)
• Level 4 41095 (34 of Formulaic)
• Stochastic reserves not materially different

46
Considerations for Application of
Principles-based Reserving to In force Contracts
47
Arguments For the Application of PBRto All In
force Contracts

• Measures the risks of a company more
  appropriately than current formulaic reserves
• Provides a consistent methodology for all
  business.
• Consistent with international actuarial and
  accounting directions
• Constitutes a more rigorous approach for all
  blocks of business
• Provides better information for regulators
• Reduces those reserves that are redundant under
  current regulation and strengthens those reserves
  that are inadequate under current regulation and
  will tend to lessen dependence on complex
  reinsurance and financing solutions

48
Arguments For the Application of PBRto All In
force Contracts

• Consistent with Enterprise Risk Management in
  that
• Incorporates risk of the entire block of business
  
• The reserve will allow some offset of covariant
  risks
• Reserves are set using the same or similar models
  to those that should be used to manage the
  business.
• 4. Allows the entire asset portfolio to be
  reflected in the reserve calculation reducing
  the subjectivity involved in allocating assets
  between PBR and non-PBR liabilities
• Mitigates the change in the pattern of margins
  under PBR as compared to the current formulaic
  approach
• Potentially reduces on-going costs by not
  requiring companies to maintain multiple reserve
  approaches  

49
Arguments Against the Application of PBR to All
In force Contracts

• Potential for large reserve discontinuity if
  inforce block is large relative to new business 
• May have significant tax implications
• Retroactive changes in reserve method are not
  permitted for tax purposes
• Reserves increase will not increase
  tax-deductible reserve but reserve decreases
  will likely decrease tax-deductible reserve
• Does not allow for as long a learning period
  with respect to the overall application systems
  and peer review before the approach is applied to
  a large block of business

50
Arguments Against the Application of PBR to All
In force Contracts

• System implications and training may initially
  lead to large implementation costs.
• Most but not all past changes in reserve
  methodology have not been applied to inforce
  business.
• Some blocks may be very small or the reserves may
  already be equal to the Cash Surrender Value
  creating a lot of additional work for little or
  no value.

51
Application of PBR to Subset of Inforce

• Three recent dates with significant changes to
  formulaic reserves
• January 1 2000 Reg XXX
• January 1 2003 AG38 section 8
• July 1 2005 revised AG38 section 8
• Inforce contracts subject to these reserve
  standards are possible subsets that could be
  subject to application of PBR
• Some of the arguments against application to
  inforce contracts are mitigated
• But many of the problems described above dealing
  with the application to inforce contracts still
  exist

52
Phased-in Application of PBR

• Another option is to phase-in the application to
  inforce contracts over time (or phase-in the
  effect over time).
• For example initially PBR would be prospective
  only and then all or a portion of inforce
  contracts would be phased on over X years.
• Some of the arguments against application to
  inforce contracts are mitigated
• But many of the problems described above dealing
  with the application to inforce contracts still
  exist

53
ACLI Interim Proposal

• ACLI has a strong commitment to see PBR
  implemented
• Concern it will take several years to get there
• Need an interim solution by 4/1/07
• Interim solution elements
• Split 2001 CSO table into preferred and residual
  tables
• Allow use of lapse assumptions for UL SG
• Allow non-premium paying UL SG contracts to use
  the surrender charge offset

54
Preferred Mortality Project

• Reasons for project
• An essential element of a principles-based
  reserving system
• Output from project may be useful as valuation
  tables under the traditional rule-based
  reserving
• Produce experience tables of greater utility to
  company actuaries

55
Project Oversight

• AAA-SOA Preferred Mortality Project Oversight
  Group (POG).
• The POG oversees the teams responsible for
  completing assigned tasks and project funding.
• POG members represent different stakeholders
• SOA (2)
• AAA (2)
• Regulators (2)
• Insurance Industry (ACLI NALC ALIA)
• Project tasks assigned to 6 teams.

56
POG Teams

• Data Validation
• Underwriting Criteria
• Experience Analysis
• Valuation Basic Tables
• Implementation
• Valuation Tables

57
Deliverables

• Valuation Basic Tables and Valuation Tables
  submitted to the NAIC March 2007
• Infrastructure for future preferred mortality
  tables

58
Challenges

• Develop the number of valuation basic tables
  (without margins) and valuation tables (with
  margins) expected by the AAA LRWG
• Lack of data for later policy duration in the
  select period
• Tables for a sufficient number of underwriting
  classes
• Persistence of the relationship between
  underwriting class mortality